Release 1.7.3

We don't make use of this feature, so it's not necessary.

Ref: af52be12fa/tiktoken-rs/Cargo.toml (L21)

.github/workflows/workflow.yml

@@ -3,13 +3,14 @@ on:
   push:
     branches: [ "main" ]
     tags: [ "v*" ]
-  schedule:
-    - cron: '0 0 * * 1'
 permissions:
   checks: write
   contents: write
   packages: write
   pull-requests: read
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: false
 jobs:
   test-and-clippy:
     name: Unit testing and linting
@@ -17,20 +18,46 @@ jobs:
     steps:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@stable
+      - name: Install SQLite3
+        run: sudo apt-get update && sudo apt-get install -y libsqlite3-dev
       - run: cargo test --all-features
       - run: cargo clippy
 
-  build-publish:
+  docker-clean-metadata:
-    name: Build and Publish
+    runs-on: ubuntu-latest
-    runs-on: self-hosted
+    outputs:
+      json: ${{ steps.meta.outputs.json }}
     steps:
-      - name: Set up QEMU
+      - name: Extract metadata (tags, labels) for Docker
-        uses: docker/setup-qemu-action@v3
+        id: meta
+        uses: docker/metadata-action@v5
         with:
-          platforms: arm64
+          images: |
-      - name: Set up Docker Buildx
+            ghcr.io/${{ github.repository }}
-        uses: docker/setup-buildx-action@v1
+          tags: |
-      - name: Login to ghcr.io
+            type=raw,value=latest,enable=${{ github.ref_name == 'main' }}
+            type=semver,pattern={{raw}}
+
+  docker-build:
+    permissions:
+      contents: read
+      packages: write
+      attestations: write
+      id-token: write
+    strategy:
+      matrix:
+        include:
+          - os: self-hosted
+            arch: amd64
+          - os: ubuntu-24.04-arm
+            arch: arm64
+
+    runs-on: ${{ matrix.os }}
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+      - name: Log in to the GitHub Container registry
         uses: docker/login-action@v3
         with:
           registry: ghcr.io
@@ -40,17 +67,41 @@ jobs:
         id: meta
         uses: docker/metadata-action@v5
         with:
-          images: |
-            ghcr.io/${{ github.repository }}
-            registry.etke.cc/${{ github.repository }}
           tags: |
             type=raw,value=latest,enable=${{ github.ref_name == 'main' }}
             type=semver,pattern={{raw}}
-      - name: Build and push
+          flavor: |
+            latest=auto
+            suffix=-${{ matrix.arch }},onlatest=true
+          images: |
+            ghcr.io/${{ github.repository }}
+
+      - name: Build and push Docker images
         uses: docker/build-push-action@v6
         with:
-          platforms: linux/amd64,linux/arm64
           push: true
           tags: ${{ steps.meta.outputs.tags }}
           labels: ${{ steps.meta.outputs.labels }}
-          file: Dockerfile.ci
+
+  docker-manifest:
+    needs:
+      - docker-build
+      - docker-clean-metadata
+    runs-on: ubuntu-latest
+
+    strategy:
+      matrix:
+        image: ${{ fromJson(needs.docker-clean-metadata.outputs.json).tags }}
+
+    steps:
+      - name: Log in to the GitHub Container registry
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Create and push manifest
+        run: |
+          docker manifest create ${{ matrix.image }} ${{ matrix.image }}-amd64 ${{ matrix.image }}-arm64
+          docker manifest push ${{ matrix.image }}

CHANGELOG.md

@@ -1,6 +1,62 @@
+# (2025-06-10) Version 1.7.3
+
+- (**Internal Improvement**) Dependency updates. This version is based on [mxlink](https://crates.io/crates/mxlink)@1.8.0 (which is based on the newly released [matrix-sdk](https://crates.io/crates/matrix-sdk)@[0.12.0](https://github.com/matrix-org/matrix-rust-sdk/releases/tag/matrix-sdk-0.12.0), which contains fixes for important security vulnerabilities)
+
+# (2025-05-11) Version 1.7.2
+
+- (**Bugfix**) Allow `image_generation.size` configuration value for OpenAI to be `null` to allow the model to choose the size automatically and default to that
+
+# (2025-05-11) Version 1.7.1
+
+- (**Bugfix**) Fix lack of documentation for the new [image-editing](./docs/features.md#-image-editing) feature in the `!bai usage` command's output
+
+# (2025-05-10) Version 1.7.0
+
+- (**Feature**) Add vision support to the OpenAI and Anthropic providers. You can now mix text and images in your conversations - fixes [issue #5](https://github.com/etkecc/baibot/issues/5)
+
+- (**Feature**) Add [image-editing](./docs/features.md#-image-editing) support to the OpenAI provider
+
+- (**Improvement**) Add compatibility with OpenAI's `gpt-image-1` model - fixes [issue #40](https://github.com/etkecc/baibot/issues/40)
+
+- (**Change**) Rework [image-creation](./docs/features.md#-image-creation) to avoid command conflicts with [image-editing](./docs/features.md#-image-editing). The image-creation command syntax is now `!bai image create <prompt>` (previously: `!bai image <prompt>`).
+
+- (**Internal Improvement**) Dependency and compiler updates
+
+> [!WARNING]
+> Unlike other releases, this release is not published to [crates.io](https://crates.io), because it relies on multiple library forks (`async-openai` and `anthropic-rs`) sourced from Github.
+
+
+# (2025-04-12) Version 1.6.0
+
+- (**Internal Improvement**) Dependency updates. This version is based on [mxlink](https://crates.io/crates/mxlink)@1.7.0 (which is based on the newly released [matrix-sdk](https://crates.io/crates/matrix-sdk)@[0.11.0](https://github.com/matrix-org/matrix-rust-sdk/releases/tag/matrix-sdk-0.11.0))
+
+
+# (2025-03-31) Version 1.5.1
+
+- (**Internal Improvement**) Dependency updates
+
+# (2025-02-27) Version 1.5.0
+
+- (**Feature**) Add support for sending Speech-to-Text replies for [Transcribe-only mode](./docs/features.md#transcribe-only-mode) as regular text messages instead of notices and doing it so by default ([a1bd292752](https://github.com/etkecc/baibot/commit/a1bd292752bdd37a196788c73d00b5619e843a78)) - improvement for [issue #14](https://github.com/etkecc/baibot/issues/14). See [🦻 Speech-to-Text / 🪄 Message Type for non-threaded only-transcribed messages](./docs/configuration/speech-to-text.md#-message-type-for-non-threaded-only-transcribed-messages) for details.
+
+- (**Feature**) Add config setting controlling if a self-introduction message is posted after joining a room ([c051da2f4a](https://github.com/etkecc/baibot/commit/c051da2f4a161de0974ebb917f7a52d01f5a001f)) - fixes [issue #32](https://github.com/etkecc/baibot/issues/32). You may wish to add a `room.post_join_self_introduction_enabled` property to your configuration. See the [sample config](./etc/app/config.yml.dist) for details. If unspecified, it defaults to `true` anyway which preserves the old behavior.
+
+- (**Feature**) Add support for configuring `max_completion_tokens` for OpenAI ([47d8edea70](https://github.com/etkecc/baibot/commit/47d8edea705a44aa25a9bfaec4888c0f9ea8700e))
+
+- (**Improvement**) Dependency updates. This version is based on [mxlink](https://crates.io/crates/mxlink)@1.6.1 (which is based on the newly released [matrix-sdk](https://crates.io/crates/matrix-sdk)@[0.10.0](https://github.com/matrix-org/matrix-rust-sdk/releases/tag/matrix-sdk-0.10.0))
+
+- (**Improvement**) Populate image/audio attachment `body` with a filename, not with text to avoid incorrect rendering in Element Web, etc. ([ec1879d212](https://github.com/etkecc/baibot/commit/ec1879d212fa8d6e5f8590486e94c72abfcb75a5))
+
+- (**Improvement**) Replace Anthropic library ([anthropic-rs](https://crates.io/crates/anthropic-rs) -> [anthropic](https://crates.io/crates/anthropic)) and switch default recommended model (`claude-3-5-sonnet-20240620` -> `claude-3-7-sonnet-20250219`) ([692d61b239](https://github.com/etkecc/baibot/commit/692d61b2398f073b81d32d4cbe8145ab3929e48c)) - fixes [issue #22](https://github.com/etkecc/baibot/issues/22)
+
+- (**Internal Improvement**) Switch to native building of `arm64` container images to decrease total build times from ~40 minutes to ~8 minutes ([6719538530b](https://github.com/etkecc/baibot/commit/6719538530bf76b3ff2d24077b2a7fa868276b79))
+
+- (**Internal Improvement**) Various other internal changes, including upgrading [Rust from 1.82 to 1.85 and switching to Rust edition 2024](https://blog.rust-lang.org/2025/02/20/Rust-1.85.0.html)
+
+
 # (2024-12-12) Version 1.4.1
 
-- (**Bugfix**) Fix detection for whether the bot is the last member in a room, to avoid incorrectly leaving multi-user rooms that have had at least one person `leave` ([d2d846816e](https://github.com/etkecc/baibot/commit/d2d846816e58c132e30cad28440552453afe1af4))
+- (**Bugfix**) Fix detection for whether the bot is the last member in a room, to avoid incorrectly leaving multi-user rooms that have had at least one person `leave` ([3c47d40781](https://github.com/etkecc/baibot/commit/3c47d407819aa9c0121117a411858238724f06da))
 
 
 # (2024-11-19) Version 1.4.0

Cargo.toml

@@ -7,32 +7,33 @@ license = "AGPL-3.0-or-later"
 readme = "README.md"
 keywords = ["matrix", "chat", "bot", "AI", "LLM"]
 include = ["/etc/assets/baibot-torso-768.png", "/src", "/README.md", "/CHANGELOG.md", "/LICENSE"]
-version = "1.4.1"
+version = "1.7.3"
-edition = "2021"
+edition = "2024"
 
 [lib]
 name = "baibot"
 path = "src/lib.rs"
 
 [dependencies]
-anthropic-rs = "0.1.*"
+anthropic = { git = "https://github.com/etkecc/anthropic-rs.git", branch = "fix-content-block-image" }
 anyhow = "1.0.*"
-async-openai = "0.26.*"
+async-openai = { git = "https://github.com/etkecc/async-openai", branch = "async-openai-v0.28.1-patched" }
 base64 = "0.22.*"
 chrono = { version = "0.4.*", default-features = false, features = ["std", "now"] }
 # We'd rather not depend on this, but we cannot use the ruma-events EventContent macro without it.
-matrix-sdk = { version = "0.8.0", default-features = false }
+# We add the `native-tls` feature, because of https://github.com/etkecc/rust-mxlink/issues/1
+matrix-sdk = { version = "0.12.0", default-features = false, features = ["native-tls"] }
 mxidwc = "1.0.*"
-mxlink = ">=1.4.1"
+mxlink = ">=1.8.0"
 etke_openai_api_rust = "0.1.*"
 quick_cache = "0.6.*"
 regex = "1.11.*"
 serde = { version = "1.0.*", features = ["derive"], default-features = false }
 serde_json = "1.0.*"
 serde_yaml = "0.9.*"
-tempfile = "3.14.*"
+tempfile = "3.20.*"
-tiktoken-rs = { version = "0.6.*", features = ["async-openai"] }
+tiktoken-rs = { version = "0.7.*", default-features = false }
-tokio = { version = "1.41.*", features = ["rt", "rt-multi-thread", "macros"] }
+tokio = { version = "1.45.*", features = ["rt", "rt-multi-thread", "macros"] }
 tracing = "0.1.*"
 tracing-subscriber = { version = "0.3.*", features = ["env-filter"] }
 url = "2.5.*"

Dockerfile

@@ -4,7 +4,7 @@
 #                                     #
 #######################################
 
-FROM docker.io/rust:1.82.0-slim-bookworm AS build
+FROM docker.io/rust:1.86.0-slim-bookworm AS build
 
 RUN apt-get update && apt-get install -y build-essential pkg-config libssl-dev libsqlite3-dev
 

Dockerfile.ci

@@ -4,7 +4,7 @@
 #                                     #
 #######################################
 
-FROM docker.io/rust:1.82.0-slim-bookworm AS build
+FROM docker.io/rust:1.86.0-slim-bookworm AS build
 
 RUN apt-get update && apt-get install -y build-essential pkg-config libssl-dev libsqlite3-dev
 

README.md

@@ -17,10 +17,10 @@ It's influenced by [chaz](https://github.com/arcuru/chaz), but does **not** use
 
 - Supports **different use purposes** (depending on the [☁️ provider](./docs/providers.md) & model):
 
-  - [💬 text-generation](./docs/features.md#-text-generation): communicating with you via text
+  - [💬 text-generation](./docs/features.md#-text-generation): communicating with you via text (though certain models may "see" images as well)
   - [🦻 speech-to-text](./docs/features.md#-speech-to-text): turning your voice messages into text
   - [🗣️ text-to-speech](./docs/features.md#%EF%B8%8F-text-to-speech): turning bot or users text messages into voice messages
-  - [🖌️ image-generation](./docs/features.md#%EF%B8%8F-image-generation): generating images based on instructions
+  - [🖌️ image-generation](./docs/features.md#image-generation): creating and editing images based on instructions
 
 - 🪄 Supports [seamless voice interaction](./docs/features.md#seamless-voice-interaction) (turning user voice messages into text, answering in text, then turning that text back into voice)
 

docs/agents.md

@@ -35,7 +35,7 @@ Depending on where the agent is defined (within a room, globally, or [statically
 
 When creating an agent, you will be given some sample [YAML](https://en.wikipedia.org/wiki/YAML) configuration which you can use to customize the agent's behavior.
 
-This configuration varies depending on the [☁️ provider](./providers.md) used and the capabilities of the agent. Based on the configuration keys you pass, certain features will be enabled or disabled. For example, if you skip the `image_generation` key for an [OpenAI](./providers.md#openai) agent, it won't be able to generate images (see [🖌️ Image Generation](./features.md#-image-generation)).
+This configuration varies depending on the [☁️ provider](./providers.md) used and the capabilities of the agent. Based on the configuration keys you pass, certain features will be enabled or disabled. For example, if you skip the `image_generation` key for an [OpenAI](./providers.md#openai) agent, it won't be able to generate images (see [🖌️ Image Creation](./features.md#-image-creation), [🎨 Image Editing](./features.md#-image-editing), [🫵 Sticker Creation](./features.md#-sticker-creation)).
 
 After making your modifications to the sample YAML, you submit it back to the bot and the new agent will be created.
 

docs/configuration/README.md

@@ -40,7 +40,7 @@ You can adjust the following settings per room and/or globally:
 - [💬 Text Generation](text-generation.md)
 - [🦻 Speech-to-Text](speech-to-text.md)
 - [🗣️ Text-to-Speech](text-to-speech.md)
-- [🖌️ Image Generation](image-generation.md)
+- [🖌️ Image Creation](image-generation.md)
 - [🤝 Handlers](handlers.md)
 
 Refer to the bot's help messages (as a response to a `!bai config` help command) for the most up-to-date information on what Room Settings can be configured.

docs/configuration/handlers.md

@@ -8,10 +8,10 @@ You can also use **different models within the same room** (e.g. [💬 text-gene
 
 The bot supports the following use-purposes:
 
-- [💬 text-generation](../features.md#-text-generation): communicating with you via text
+- [💬 text-generation](../features.md#-text-generation): communicating with you via text (though certain models may "see" images as well)
 - [🦻 speech-to-text](../features.md#-speech-to-text): turning your voice messages into text
 - [🗣️ text-to-speech](../features.md#️-text-to-speech): turning bot or users text messages into voice messages
-- [🖌️ image-generation](../features.md#-image-generation): generating images based on instructions
+- [🖌️ image-generation](../features.md#image-generation): generating images based on instructions
 
 In a given room, each different purpose can be served by a different [provider](../providers.md) and model. This combination of provider and model configuration is called an [🤖 agent](../agents.md). Each purpose can be served by a different **handler** agent.
 

docs/configuration/image-generation.md

@@ -1,9 +1,11 @@
 
-## 🖌️ Image Generation
+## Image Generation
 
-The Image Generation feature is not configurable at this moment.
+The Image Creation and Image Editing features are not configurable at this moment.
 
 You may also wish to see:
 
-- [🌟 Features / 🖌️ Image Generation](../features.md#-image-generation) for a higher-level introduction to the Image Generation features
+- [🌟 Features / Image Generation / 🖌️ Image Creation](../features.md#-image-creation) for a higher-level introduction to the Image Creation features
-- [📖 Usage / 🖌️ Image Generation](../usage.md#-image-generation) section for more details on how to use the bot for Image Generation in a room
+- [🌟 Features / Image Generation / 🎨 Image Editing](../features.md#-image-editing) for a higher-level introduction to the Image Editing features
+- [📖 Usage / Image Generation / 🖌️ Creating Images](../usage.md#-creating-images) section for more details on how to use the bot for Image Creation in a room
+- [📖 Usage / Image Generation / 🎨 Editing images](../usage.md#-editing-images) section for more details on how to use the bot for Image Editing in a room

docs/configuration/speech-to-text.md

@@ -23,6 +23,19 @@ The following configuration values are recognized:
 Example: `!bai config room speech-to-text set-flow-type ignore` (this can also be set globally, see [🛠️ Room Settings](./README.md#room-settings))
 
 
+### 🪄 Message Type for non-threaded only-transcribed messages
+
+Controls how the transcribed text of voice messages is sent to the chat when Flow Type = `only_transcribe`.
+
+The following configuration values are recognized:
+
+- (default) `text`: the transcribed text is sent as a regular message. This is more convenient if you'd like to forward the transcribed message to other rooms.
+
+- `notice`: the transcribed text is sent as a notice message. This provides better compatibility with other bots in the room, as they are less likely to interact with messages of type notice.
+
+Example: `!bai config room speech-to-text set-msg-type-for-non-threaded-only-transcribed-messages notice` (this can also be set globally, see [🛠️ Room Settings](./README.md#room-settings))
+
+
 ### 🔤 Language
 
 Lets you specify the language of the input voice messages, to avoid using auto-detection.

docs/development.md

@@ -93,7 +93,7 @@ For getting started most quickly (and locally), we recommend using [LocalAI](#lo
 
 **Ollama is most lightweight** (~2GB for the container image + ~1.6GB for the model), but supports only [💬 text-generation](./features.md#-text-generation).
 
-**LocalAI requires 4x more disk space** (~6GB for the container image + ~12GB for the models), but supports [💬 text-generation](./features.md#-text-generation), [🗣️ text-to-speech](./features.md#️-text-to-speech), [🦻 speech-to-text](./features.md#-speech-to-text) and [🖼️ image-generation](./features.md#️-image-generation).
+**LocalAI requires 4x more disk space** (~6GB for the container image + ~12GB for the models), but supports [💬 text-generation](./features.md#-text-generation), [🗣️ text-to-speech](./features.md#️-text-to-speech), [🦻 speech-to-text](./features.md#-speech-to-text) and [🖼️ image-generation](./features.md#️-image-creation).
 
 **OpenAI supports all of these capabilities** as well and does not require powerful hardware or lots of disk space. However, it requires signup and an API key.
 

docs/features.md

@@ -8,7 +8,7 @@ You can also use **different models within the same room** (e.g. [💬 text-gene
 
 The bot supports the following use-purposes:
 
-- [💬 text-generation](#-text-generation): communicating with you via text
+- [💬 text-generation](#-text-generation): communicating with you via text (though certain models may "see" images as well)
 - [🦻 speech-to-text](#-speech-to-text): turning your voice messages into text
 - [🗣️ text-to-speech](#%EF%B8%8F-text-to-speech): turning bot or users text messages into voice messages
 - [🖌️ image-generation](#%EF%B8%8F-image-generation): generating images based on instructions
@@ -22,10 +22,12 @@ For more information about configuring handlers, see the [🤝 Handlers / Config
 
 ### 💬 Text Generation
 
-Text Generation is the bot's ability to **respond to users' text messages with text**.
+Text Generation is the bot's ability to **respond to users' messages with text**.
 
 ![Screenshot of Text Generation - a user sends a message and the bot replies in a new conversation thread](./screenshots/text-generation.webp)
 
+Some models also support vision, so you may be able to mix text and images in the same conversation.
+
 In multi-user (group) rooms, to avoid disturbing the normal conversation between people, the bot is auto-configured to only respond to messages starting with the command prefix (`!bai`) or direct mentions via the [💬 Text Generation / 🗟 Prefix Requirement Type](./configuration/text-generation.md#-prefix-requirement-type) setting.
 
 Normally, the bot only responds to allowed [👥 Users](./access.md#-users). In certain cases, it's useful for an allowed user to provoke the bot to respond even in foreign threads or reply chains. You can learn more about this feature in the [On-demand involvement](./features.md#on-demand-involvement) section below.
@@ -136,27 +138,45 @@ To operate in this mode, you can:
 
 - adjust the [🦻 Speech-to-Text / 🪄 Flow Type](./configuration/speech-to-text.md#-flow-type) setting to make the bot only transcribe (without doing [💬 Text Generation](#-text-generation)): `!bai config room speech-to-text set-flow-type only_transcribe`
 
+- optionally adjust [🦻 Speech-to-Text / 🪄 Message Type for non-threaded only-transcribed messages](./configuration/speech-to-text.md#-message-type-for-non-threaded-only-transcribed-messages), if you'd like to bot to send messages of type `notice` (for better compatibility with other bots in the room) instead of sending regular `text` messages (default)
 
-### 🖌️ Image Generation
 
-Image generation is the bot's ability to **generate images** based on text prompts.
+### Image Generation
 
-See a [🖼️ Screenshot of the Image Generation feature](./screenshots/image-generation.webp).
+#### 🖌️ Image Creation
+
+Image creation is the bot's ability to **create images** based on text prompts.
+
+See a [🖼️ Screenshot of the Image Creation feature](./screenshots/image-creation.webp).
 
 You may also wish to see:
 
 - [🛠️ Configuration / 🖌️ Image Generation](./configuration/image-generation.md) for configuration options related to Image Generation
-- [📖 Usage / 🖌️ Image Generation](./usage.md#-image-generation) section for more details on how to use the bot for Image Generation in a room
+- [📖 Usage / Image Generation / 🖌️ Creating Images](./usage.md#-creating-images) section for more details on how to use the bot for Image Creation in a room
-- [🫵 Sticker Generation](#-sticker-generation) - a special case of Image Generation
+- [🖌️ Image Editing](#️-image-editing) - another image generation feature
+- [🫵 Sticker Creation](#-sticker-creation) - a special case of Image Creation
 
 
-### 🫵 Sticker Generation
+#### 🎨 Image Editing
 
-Sticker generation is the bot's ability to **generate sticker** images based on text prompts. It's a special case of [🖌️ Image Generation](#️-image-generation).
+Image editing is the bot's ability to **edit images** based on a prompt and one or more existing images.
 
-See a [🖼️ Screenshot of the Sticker Generation feature](./screenshots/sticker-generation.webp).
+See a [🖼️ Screenshot of the Image Editing feature (manipulating a single image)](./screenshots/image-editing-single-image.webp) and a [🖼️ Screenshot of the Image Editing feature (manipulating multiple images)](./screenshots/image-editing-multiple-images.webp).
 
-See [📖 Usage / 🖌️ Image Generation / Generating Stickers](./usage.md#generating-stickers) for details.
+You may also wish to see:
+
+- [🛠️ Configuration / 🖌️ Image Generation](./configuration/image-generation.md) for configuration options related to Image Generation
+- [📖 Usage / Image Generation / 🎨 Editing images](./usage.md#-editing-images) section for more details on how to use the bot for Image Editing in a room
+- [🖌️ Image Creation](#️-image-creation) - another image generation feature
+
+
+#### 🫵 Sticker Creation
+
+Sticker generation is the bot's ability to **generate sticker** images based on text prompts. It's a special case of [🖌️ Image Creation](#️-image-creation).
+
+See a [🖼️ Screenshot of the Sticker Creation feature](./screenshots/sticker-generation.webp).
+
+See [📖 Usage / Image Generation / 🫵 Creating Stickers](./usage.md#-creating-stickers) for details.
 
 
 ### 🔒 Encryption

docs/installation.md

@@ -53,6 +53,7 @@ CONTAINER_IMAGE_NAME=ghcr.io/etkecc/baibot:v1.0.0
   --env BAIBOT_PERSISTENCE_DATA_DIR_PATH=/data \
   --mount type=bind,src=/path/to/config.yml,dst=/app/config.yml,ro \
   --mount type=bind,src=/path/to/data,dst=/data \
+  --tmpfs=/tmp:rw,noexec,nosuid,size=1024m \
   $CONTAINER_IMAGE_NAME
 ```
 

docs/providers.md

@@ -23,7 +23,7 @@ The list of supported providers is below.
 
 ### How to choose a provider
 
-If you're not sure which provider to start with, **we recommend [OpenAI](#openai)** as it's the most popular and has the **widest range of capabilities**: [💬 text-generation](./features.md#-text-generation), [🖌️ image-generation](./features.md#️-image-generation), [🦻 speech-to-text](./features.md#-speech-to-text), [🗣️ text-to-speech](./features.md#️-text-to-speech).
+If you're not sure which provider to start with, **we recommend [OpenAI](#openai)** as it's the most popular and has the **widest range of capabilities**: [💬 text-generation](./features.md#-text-generation) (no vision), [🖌️ image-generation](./features.md#️image-generation), [🦻 speech-to-text](./features.md#-speech-to-text), [🗣️ text-to-speech](./features.md#️-text-to-speech).
 
 You don't need to choose just one though. The bot supports [mixing & matching models](./features.md#-mixing--matching-models), so you can use multiple providers at the same time.
 
@@ -47,7 +47,7 @@ You don't need to choose just one though. The bot supports [mixing & matching mo
 
 - 🆔 Identifier: `anthropic`
 - 🔗 Links: [🏠 Home page](https://www.anthropic.com/), [🌐 Wiki](https://en.wikipedia.org/wiki/Anthropic), [👤 Sign up](https://console.anthropic.com/), [📋 Models list](https://docs.anthropic.com/en/docs/about-claude/models)
-- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (incl. vision)
 - 🗲 Quick start:
   - create a room-local agent: `!bai agent create-room-local anthropic my-anthropic-agent`
   - create a global agent: `!bai agent create-global anthropic my-anthropic-agent`
@@ -61,7 +61,7 @@ You don't need to choose just one though. The bot supports [mixing & matching mo
 
 - 🆔 Identifier: `groq`
 - 🔗 Links: [🏠 Home page](https://groq.com/), [🌐 Wiki](https://en.wikipedia.org/wiki/Groq), [👤 Sign up](https://console.groq.com/login), [📋 Models list](https://console.groq.com/docs/models)
-- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation), [🦻 speech-to-text](./features.md#-speech-to-text)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (no vision), [🦻 speech-to-text](./features.md#-speech-to-text)
 - 🗲 Quick start:
   - create a room-local agent: `!bai agent create-room-local groq my-groq-agent`
   - create a global agent: `!bai agent create-global groq my-groq-agent`
@@ -75,7 +75,7 @@ You don't need to choose just one though. The bot supports [mixing & matching mo
 
 - 🆔 Identifier: `localai`
 - 🔗 Links: [🏠 Home page](https://localai.io/), [📋 Models list](https://localai.io/gallery.html)
-- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation), [🗣️ text-to-speech](./features.md#️-text-to-speech), [🦻 speech-to-text](./features.md#-speech-to-text)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (no vision), [🗣️ text-to-speech](./features.md#️-text-to-speech), [🦻 speech-to-text](./features.md#-speech-to-text)
 - 🗲 Quick start:
   - create a room-local agent: `!bai agent create-room-local localai my-localai-agent`
   - create a global agent: `!bai agent create-global localai my-localai-agent`
@@ -89,7 +89,7 @@ You don't need to choose just one though. The bot supports [mixing & matching mo
 
 - 🆔 Identifier: `mistral`
 - 🔗 Links: [🏠 Home page](https://mistral.ai/), [🌐 Wiki](https://en.wikipedia.org/wiki/Mistral_AI), [👤 Sign up](https://auth.mistral.ai/ui/registration), [📋 Models list](https://docs.mistral.ai/getting-started/models/)
-- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (no vision)
 - 🗲 Quick start:
   - create a room-local agent: `!bai agent create-room-local mistral my-mistral-agent`
   - create a global agent: `!bai agent create-global mistral my-mistral-agent`
@@ -103,7 +103,7 @@ You don't need to choose just one though. The bot supports [mixing & matching mo
 
 - 🆔 Identifier: `ollama`
 - 🔗 Links: [🏠 Home page](https://ollama.com/), [📋 Models list](https://ollama.com/library)
-- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (no vision)
 - 🗲 Quick start:
   - create a room-local agent: `!bai agent create-room-local ollama my-ollama-agent`
   - create a global agent: `!bai agent create-global ollama my-ollama-agent`
@@ -120,7 +120,7 @@ For services which are not fully compatible with the OpenAI API, consider using
 
 - 🆔 Identifier: `openai`
 - 🔗 Links: [🏠 Home page](https://openai.com/), [🌐 Wiki](https://en.wikipedia.org/wiki/OpenAI), [👤 Sign up](https://platform.openai.com/signup), [📋 Models list](https://platform.openai.com/docs/models)
-- 🌟 Capabilities: [🖌️ image-generation](./features.md#️-image-generation), [💬 text-generation](./features.md#-text-generation), [🗣️ text-to-speech](./features.md#️-text-to-speech), [🦻 speech-to-text](./features.md#-speech-to-text)
+- 🌟 Capabilities: [🖌️ image-generation](./features.md#️-image-creation), [💬 text-generation](./features.md#-text-generation) (incl. vision), [🗣️ text-to-speech](./features.md#️-text-to-speech), [🦻 speech-to-text](./features.md#-speech-to-text)
 - 🗲 Quick start:
   - create a room-local agent: `!bai agent create-room-local openai my-openai-agent`
   - create a global agent: `!bai agent create-global openai my-openai-agent`
@@ -140,7 +140,7 @@ Some of these popular services already have **shortcut** providers (leading to t
 This provider is just as featureful as the [OpenAI](#openai) provider, but is more compatible with services which do not fully adhere to the [OpenAI API spec](https://github.com/openai/openai-openapi/).
 
 - 🆔 Identifier: `openai-compatible`
-- 🌟 Capabilities: [🖌️ image-generation](./features.md#️-image-generation), [💬 text-generation](./features.md#-text-generation), [🗣️ text-to-speech](./features.md#️-text-to-speech), [🦻 speech-to-text](./features.md#-speech-to-text)
+- 🌟 Capabilities: [🖌️ image-generation](./features.md#️-image-creation), [💬 text-generation](./features.md#-text-generation) (no vision), [🗣️ text-to-speech](./features.md#️-text-to-speech), [🦻 speech-to-text](./features.md#-speech-to-text)
 - 🗲 Quick start:
   - create a room-local agent: `!bai agent create-room-local openai-compatible my-openai-compatible-agent`
   - create a global agent: `!bai agent create-global openai-compatible my-openai-compatible-agent`
@@ -154,7 +154,7 @@ This provider is just as featureful as the [OpenAI](#openai) provider, but is mo
 
 - 🆔 Identifier: `openrouter`
 - 🔗 Links: [🏠 Home page](https://openrouter.ai/), [👤 Sign up](https://openrouter.ai/), [📋 Models list](https://openrouter.ai/models)
-- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (no vision)
 - 🗲 Quick start:
   - create a room-local agent: `!bai agent create-room-local openrouter my-openrouter-agent`
   - create a global agent: `!bai agent create-global openrouter my-openrouter-agent`
@@ -168,7 +168,7 @@ This provider is just as featureful as the [OpenAI](#openai) provider, but is mo
 
 - 🆔 Identifier: `together-ai`
 - 🔗 Links: [🏠 Home page](https://www.together.ai/), [👤 Sign up](https://api.together.ai/signup), [📋 Models list](https://api.together.xyz/models)
-- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (no vision)
 - 🗲 Quick start:
   - create a room-local agent: `!bai agent create-room-local together-ai my-together-ai-agent`
   - create a global agent: `!bai agent create-global together-ai my-together-ai-agent`

docs/sample-provider-configs/anthropic.yml

@@ -1,7 +1,7 @@
 base_url: https://api.anthropic.com/v1
 api_key: YOUR_API_KEY_HERE
 text_generation:
-  model_id: claude-3-5-sonnet-20240620
+  model_id: claude-3-7-sonnet-20250219
   prompt: "You are a brief, but helpful bot called {{ baibot_name }} powered by the {{ baibot_model_id }} model. The date/time of this conversation's start is: {{ baibot_conversation_start_time_utc }}."
   temperature: 1.0
   max_response_tokens: 8192

docs/sample-provider-configs/openai-o1.yml

@@ -6,9 +6,9 @@ text_generation:
   prompt: null
   temperature: 1.0
   # o1 models do not support max_response_tokens.
-  # They use `max_completion_tokens` as an alternative,
+  # They use `max_completion_tokens` as an alternative
-  # but we don't support it yet (see https://github.com/64bit/async-openai/issues/272).
   max_response_tokens: null
+  max_completion_tokens: 16384
   max_context_tokens: 128000
 speech_to_text:
   model_id: whisper-1
@@ -18,7 +18,7 @@ text_to_speech:
   speed: 1.0
   response_format: opus
 image_generation:
-  model_id: dall-e-3
+  model_id: gpt-image-1
-  style: vivid
+  style: null
-  size: 1024x1024
+  size: null
-  quality: standard
+  quality: null

docs/sample-provider-configs/openai.yml

@@ -1,7 +1,7 @@
 base_url: https://api.openai.com/v1
 api_key: YOUR_API_KEY_HERE
 text_generation:
-  model_id: gpt-4o
+  model_id: gpt-4.1
   prompt: "You are a brief, but helpful bot called {{ baibot_name }} powered by the {{ baibot_model_id }} model. The date/time of this conversation's start is: {{ baibot_conversation_start_time_utc }}."
   temperature: 1.0
   max_response_tokens: 16384
@@ -14,7 +14,7 @@ text_to_speech:
   speed: 1.0
   response_format: opus
 image_generation:
-  model_id: dall-e-3
+  model_id: gpt-image-1
-  style: vivid
+  style: null
-  size: 1024x1024
+  size: null
-  quality: standard
+  quality: null

docs/usage.md

@@ -11,6 +11,8 @@ This is related to the [💬 Text Generation](./features.md#-text-generation) fe
 
 If there's a text-generation handler agent configured, the bot **may** respond to messages sent in the room.
 
+Some models also support vision, so you may be able to mix text and images in the same conversation.
+
 See screenshots of:
 
 - 🖼️ [the default Text Generation flow](./screenshots/text-generation.webp) in 1:1 rooms
@@ -64,34 +66,48 @@ The speech-to-text feature triggers automatically by default, but can be adjuste
 If all your messages are in the same language, you can improve accuracy & latency by configuring the language (see [🦻 Speech-to-Text / 🔤 Language](./configuration/speech-to-text.md#-language)).
 
 
-### 🖌️ Image Generation
+### Image Generation
-
-This is related to the [🖌️ Image Generation](./features.md#️-image-generation) feature.
 
 This feature is not configurable at the moment. The configuration (size, quality, style) specified at the [🤖 agent](./agents.md) level will be used.
 
+Capabilities depend on the [☁️ provider](./providers.md) and model used.
 
-#### Generating images
 
-Simply send a command like `!bai image A beautiful sunset over the ocean` and the bot will start a threaded conversation and post an image based on your prompt.
+#### 🖌️ Creating images
 
-See a [🖼️ Screenshot of the Image Generation feature](./screenshots/image-generation.webp).
+Simply send a command like `!bai image create A beautiful sunset over the ocean` and the bot will start a threaded conversation and post an image based on your prompt.
 
-You can then, respond in the same message thread with:
+See a [🖼️ Screenshot of the Image Creation feature](./screenshots/image-creation.webp).
+
+You can then respond in the same message thread with:
 
 - more messages, to add more criteria to your prompt.
 - a message saying `again`, to generate one more image with the current prompt.
 
 
-#### Generating stickers
+#### 🎨 Editing images
 
-A variation of [generating images](#generating-images) is to generate "sticker images".
+Simply send a command like `!bai image edit Turn the following image into an anime-style drawing` and the bot will start a threaded conversation asking for more details.
 
-See a [🖼️ Screenshot of the Sticker Generation feature](./screenshots/sticker-generation.webp).
+See a [🖼️ Screenshot of the Image Editing feature (manipulating a single image)](./screenshots/image-editing-single-image.webp) and a [🖼️ Screenshot of the Image Editing feature (manipulating multiple images)](./screenshots/image-editing-multiple-images.webp).
 
-To generate a sticker, send a command like `!bai sticker A huge ramen bowl with lots of chashu and a mountain of beansprouts on top`.
+You can then respond in the same message thread with:
 
-The difference from [generating images](#generating-images) is that the bot will:
+- more messages, to add more criteria to your prompt.
+- one or more images, to provide the images that the bot will operate on.
+- a message saying `go`, to start the image generation process.
+- a message saying `again`, to prompt the bot to generate one more image edit with the current prompt.
+
+
+#### 🫵 Creating stickers
+
+A variation of [creating images](#creating-images) is creating "sticker images".
+
+See a [🖼️ Screenshot of the Sticker Creation feature](./screenshots/sticker-generation.webp).
+
+To create a sticker, send a command like `!bai sticker A huge ramen bowl with lots of chashu and a mountain of beansprouts on top`.
+
+The difference from [creating images](#creating-images) is that the bot will:
 
 - generate a smaller-resolution image (currently hardcoded to `256x256`) - smaller/quicker, but still good enough for a sticker
 - potentially switch to a different (cheaper or otherwise more suitable) model, if available

etc/app/config.yml.dist

@@ -32,6 +32,10 @@ user:
 # Command prefix. Leave empty to use the default (!bai).
 command_prefix: "!bai"
 
+room:
+  # Whether the bot should send an introduction message after joining a room.
+  post_join_self_introduction_enabled: true
+
 access:
   # Space-separated list of MXID patterns which specify who is an admin.
   admin_patterns:
@@ -72,10 +76,12 @@ agents:
     #     base_url: https://api.openai.com/v1
     #     api_key: ""
     #     text_generation:
-    #       model_id: gpt-4o
+    #       model_id: gpt-4.1
     #       prompt: "You are a brief, but helpful bot called {{ baibot_name }} powered by the {{ baibot_model_id }} model. The date/time of this conversation's start is: {{ baibot_conversation_start_time_utc }}."
     #       temperature: 1.0
     #       max_response_tokens: 16384
+    #       # Reasoning models need to use `max_completion_tokens` instead of `max_response_tokens`.
+    #       max_completion_tokens: ~
     #       max_context_tokens: 128000
     #     speech_to_text:
     #       model_id: whisper-1
@@ -85,10 +91,10 @@ agents:
     #       speed: 1.0
     #       response_format: opus
     #     image_generation:
-    #       model_id: dall-e-3
+    #       model_id: gpt-image-1
-    #       style: vivid
+    #       style: null
-    #       size: 1024x1024
+    #       size: null
-    #       quality: standard
+    #       quality: null
     #
     # - id: localai
     #   provider: localai

etc/services/core/compose.yml

@@ -1,6 +1,6 @@
 services:
   postgres:
-    image: docker.io/postgres:16.4-alpine
+    image: docker.io/postgres:16.8-alpine
     user: ${UID}:${GID}
     restart: unless-stopped
     environment:
@@ -13,7 +13,7 @@ services:
     - /etc/passwd:/etc/passwd:ro
 
   synapse:
-    image: ghcr.io/element-hq/synapse:v1.118.0
+    image: ghcr.io/element-hq/synapse:v1.129.0
     user: "${UID}:${GID}"
     restart: unless-stopped
     entrypoint: python
@@ -26,14 +26,20 @@ services:
     - ./synapse/media-store:/media-store
 
   element-web:
-    image: docker.io/vectorim/element-web:v1.11.84
+    image: ghcr.io/element-hq/element-web:v1.11.100
     user: "${UID}:${GID}"
     restart: unless-stopped
+    environment:
+      ELEMENT_WEB_PORT: 8080
     ports:
       - "${SERVICE_ELEMENT_WEB_BIND_PORT_HTTP}:8080"
     volumes:
-    - ../../etc/services/core/element-web/nginx.conf:/etc/nginx/nginx.conf:ro
     - ../../etc/services/core/element-web/config.json:/app/config.json:ro
+    tmpfs:
+      - /var/cache/nginx:rw,mode=777
+      - /var/run:rw,mode=777
+      - /tmp/element-web-config:rw,mode=777
+      - /etc/nginx/conf.d:rw,mode=777
 
 networks:
   default:

etc/services/core/element-web/config.json

@@ -3,7 +3,7 @@
     "default_is_url": "https://vector.im",
     "integrations_ui_url": "https://scalar.vector.im/",
     "integrations_rest_url": "https://scalar.vector.im/api",
-    "bug_report_endpoint_url": "https://riot.im/bugreports/submit",
+    "bug_report_endpoint_url": "https://element.io/bugreports/submit",
     "enableLabs": true,
     "roomDirectory": {
         "servers": [

etc/services/core/element-web/nginx.conf

@@ -1,60 +0,0 @@
-# This is a custom nginx configuration file that we use in the container (instead of the default one),
-# because it allows us to run nginx with a non-root user.
-#
-# For this to work, the default vhost file (`/etc/nginx/conf.d/default.conf`) also needs to be removed.
-# (mounting `/dev/null` over `/etc/nginx/conf.d/default.conf` works well)
-#
-# The following changes have been done compared to a default nginx configuration file:
-# - default server port is changed (80 -> 8080), so that a non-root user can bind it
-# - various temp paths are changed to `/tmp`, so that a non-root user can write to them
-# - the `user` directive was removed, as we don't want nginx to switch users
-
-worker_processes  1;
-
-error_log /var/log/nginx/error.log warn;
-pid /tmp/nginx.pid;
-
-
-events {
-	worker_connections  1024;
-}
-
-
-http {
-	client_body_temp_path /tmp/client_body_temp;
-	proxy_temp_path /tmp/proxy_temp;
-	fastcgi_temp_path /tmp/fastcgi_temp;
-	uwsgi_temp_path /tmp/uwsgi_temp;
-	scgi_temp_path /tmp/scgi_temp;
-
-	include       /etc/nginx/mime.types;
-	default_type  application/octet-stream;
-
-	log_format  main  '$remote_addr - $remote_user [$time_local] "$request" '
-					  '$status $body_bytes_sent "$http_referer" '
-					  '"$http_user_agent" "$http_x_forwarded_for"';
-
-	access_log  /var/log/nginx/access.log  main;
-
-	sendfile        on;
-	#tcp_nopush     on;
-
-	keepalive_timeout  65;
-
-	#gzip  on;
-
-	server {
-		listen       8080;
-		server_name  localhost;
-
-		location / {
-			root   /usr/share/nginx/html;
-			index  index.html index.htm;
-		}
-
-		error_page   500 502 503 504  /50x.html;
-		location = /50x.html {
-			root   /usr/share/nginx/html;
-		}
-	}
-}

etc/services/ollama/compose.yml

@@ -1,6 +1,6 @@
 services:
   ollama:
-    image: docker.io/ollama/ollama:0.4.1
+    image: docker.io/ollama/ollama:0.6.8
     restart: unless-stopped
     ports:
       - "${SERVICE_OLLAMA_BIND_PORT_HTTP}:11434"

justfile

@@ -32,6 +32,10 @@ run-in-container *extra_args: app-container-prepare build-container-image-debug
 test *extra_args:
 	RUST_BACKTRACE=1 cargo test {{ extra_args }}
 
+# Formats the code
+fmt:
+	RUST_BACKTRACE=1 cargo fmt --all
+
 # Builds a debug binary (target/debug/*)
 build-debug *extra_args:
 	RUST_BACKTRACE=1 cargo build {{ extra_args }}

src/agent/instantiation.rs

@@ -1,6 +1,6 @@
 use super::{
-    provider::{self, ControllerType},
     AgentDefinition, AgentProvider, PublicIdentifier,
+    provider::{self, ControllerType},
 };
 
 // Dead-code is allowed. We do not use these enum struct payloads directly,

src/agent/manager.rs

@@ -1,7 +1,7 @@
-use super::instantiation;
-use super::instantiation::AgentInstance;
 use super::AgentDefinition;
 use super::PublicIdentifier;
+use super::instantiation;
+use super::instantiation::AgentInstance;
 use crate::entity::RoomConfigContext;
 
 #[derive(Debug)]

src/agent/mod.rs

@@ -11,11 +11,11 @@ pub use manager::Manager;
 
 pub use definition::AgentDefinition;
 
-pub use instantiation::create_from_provider_and_yaml_value_config;
-pub use instantiation::default_config_for_provider;
 pub use instantiation::AgentInstance;
 pub use instantiation::Error as AgentInstantiationError;
 pub use instantiation::Result as AgentInstantiationResult;
+pub use instantiation::create_from_provider_and_yaml_value_config;
+pub use instantiation::default_config_for_provider;
 
 pub use provider::{AgentProvider, AgentProviderInfo, ControllerTrait};
 pub use purpose::AgentPurpose;

src/agent/provider/anthropic/config.rs

@@ -1,7 +1,5 @@
 use serde::{Deserialize, Serialize};
 
-use anthropic_rs::models::claude::ClaudeModel;
-
 use crate::agent::{default_prompt, provider::ConfigTrait};
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -28,6 +26,9 @@ impl ConfigTrait for Config {
         if self.base_url.is_empty() {
             return Err("The base URL must not be empty.".to_owned());
         }
+        if !self.base_url.ends_with("/v1") {
+            return Err("The base URL must end with '/v1'.".to_owned());
+        }
         if self.api_key.is_empty() {
             return Err("The API key must not be empty.".to_owned());
         }
@@ -67,5 +68,5 @@ impl Default for TextGenerationConfig {
 }
 
 fn default_text_model_id() -> String {
-    ClaudeModel::Claude35Sonnet.as_str().to_owned()
+    "claude-3-7-sonnet-20250219".to_owned()
 }

src/agent/provider/anthropic/controller.rs

@@ -1,30 +1,28 @@
 use std::fmt::Debug;
-use std::str::FromStr;
 use std::sync::Arc;
 
-use anthropic_rs::completion::message::{ContentType, System};
+use anthropic::client::{Client, ClientBuilder};
-use anthropic_rs::{
+use anthropic::types::ContentBlock;
-    client::Client as AnthropicClient, config::Config as AnthropicConfig,
-    models::claude::ClaudeModel,
-};
 
 use super::super::ControllerTrait;
-use crate::agent::provider::entity::{
-    ImageGenerationResult, PingResult, TextGenerationParams, TextGenerationResult,
-    TextToSpeechParams, TextToSpeechResult,
-};
-use crate::agent::provider::{ImageGenerationParams, SpeechToTextParams, SpeechToTextResult};
 use crate::agent::AgentPurpose;
+use crate::agent::provider::entity::{
+    ImageEditResult, ImageGenerationResult, ImageSource, PingResult, TextGenerationParams,
+    TextGenerationResult, TextToSpeechParams, TextToSpeechResult,
+};
+use crate::agent::provider::{
+    ImageEditParams, ImageGenerationParams, SpeechToTextParams, SpeechToTextResult,
+};
 use crate::conversation::llm::{
-    shorten_messages_list_to_context_size, Author as LLMAuthor, Conversation as LLMConversation,
+    Author as LLMAuthor, Conversation as LLMConversation, Message as LLMMessage,
-    Message as LLMMessage,
+    MessageContent as LLMMessageContent, shorten_messages_list_to_context_size,
 };
 use crate::strings;
 
 use super::config::Config;
 
 struct ControllerInner {
-    client: AnthropicClient,
+    client: Client,
 }
 
 #[derive(Clone)]
@@ -43,18 +41,20 @@ impl Debug for Controller {
 
 impl Controller {
     pub fn new(config: Config) -> anyhow::Result<Self> {
-        let anthropic_config =
+        // The previous library that we used expected a base URL that ends with "/v1"
-            AnthropicConfig::new(config.api_key.clone()).with_base_url(config.base_url.clone());
+        // (e.g. "https://api.anthropic.com/v1"), while the new one doesn't.
+        //
+        // To keep backward compatibility, we don't ask people to change their configuration
+        // and rather adapt by removing the "/v1" from the base URL.
+        if !config.base_url.ends_with("/v1") {
+            return Err(anyhow::anyhow!("base_url must end with '/v1'"));
+        }
 
-        let client = match AnthropicClient::new(anthropic_config) {
+        let base_url = &config.base_url[..config.base_url.len() - 3];
-            Ok(client) => client,
+        let client = ClientBuilder::default()
-            Err(err) => {
+            .api_base(base_url.to_string())
-                return Err(anyhow::anyhow!(
+            .api_key(config.api_key.clone())
-                    "Failed to create Anthropic client: {}",
+            .build()?;
-                    err.to_string()
-                ));
-            }
-        };
 
         Ok(Self {
             config,
@@ -71,7 +71,7 @@ impl ControllerTrait for Controller {
 
         let messages = vec![LLMMessage {
             author: LLMAuthor::User,
-            message_text: "Hello!".to_string(),
+            content: LLMMessageContent::Text("Hello!".to_string()),
             timestamp: chrono::Utc::now(),
         }];
 
@@ -108,7 +108,7 @@ impl ControllerTrait for Controller {
         } else {
             Some(LLMMessage {
                 author: LLMAuthor::Prompt,
-                message_text: prompt_text,
+                content: LLMMessageContent::Text(prompt_text),
                 timestamp: chrono::Utc::now(),
             })
         };
@@ -142,29 +142,19 @@ impl ControllerTrait for Controller {
 
         let mut request = super::utils::create_anthropic_message_request(conversation_messages);
 
-        let model = match ClaudeModel::from_str(&text_generation_config.model_id) {
-            Ok(model) => model,
-            Err(err) => {
-                tracing::debug!(?err, "Failed to parse model ID");
-
-                return Err(anyhow::anyhow!(
-                    "Failed to parse model ID: {}",
-                    &text_generation_config.model_id
-                ));
-            }
-        };
-
         let temperature = params
             .temperature_override
             .unwrap_or(text_generation_config.temperature);
 
         if let Some(prompt_message) = prompt_message {
-            request.system = Some(System::Text(prompt_message.message_text));
+            if let LLMMessageContent::Text(text) = &prompt_message.content {
+                request.system = text.clone();
+            }
         }
 
-        request.model = model;
+        request.model = text_generation_config.model_id.clone();
-        request.temperature = Some(temperature);
+        request.temperature = Some(temperature as f64);
-        request.max_tokens = text_generation_config.max_response_tokens;
+        request.max_tokens = text_generation_config.max_response_tokens as usize;
 
         if let Ok(request_as_json) = serde_json::to_string(&request) {
             tracing::trace!(
@@ -175,19 +165,20 @@ impl ControllerTrait for Controller {
             );
         }
 
-        let response = self.inner.client.create_message(request).await?;
+        let response = self.inner.client.messages(request).await?;
 
         tracing::trace!(?response, "Got response from Anthropic create message API");
 
         // response.content usually contains a single element, but we support handling multiple to account for all possibilities
         let mut text_parts = vec![];
         for content in response.content {
-            let content_type = content.content_type;
+            match content {
-
+                ContentBlock::Text { text } => {
-            match content_type {
+                    text_parts.push(text);
-                ContentType::Text => {
+                }
-                    text_parts.push(content.text);
+                ContentBlock::Image { .. } => {
-                } // There are no other content types to handle yet, but there may be in the future
+                    text_parts.push("The model responded with an image".to_string());
+                }
             }
         }
 
@@ -219,6 +210,15 @@ impl ControllerTrait for Controller {
         Err(anyhow::anyhow!("Image generation not supported"))
     }
 
+    async fn create_image_edit(
+        &self,
+        _prompt: &str,
+        _images: Vec<ImageSource>,
+        _params: ImageEditParams,
+    ) -> anyhow::Result<ImageEditResult> {
+        Err(anyhow::anyhow!("Image editing is not supported"))
+    }
+
     async fn text_to_speech(
         &self,
         _input: &str,

src/agent/provider/anthropic/mod.rs

@@ -7,8 +7,8 @@ pub use controller::Controller;
 
 use super::super::AgentInstantiationError;
 use super::super::AgentInstantiationResult;
-use super::controller::ControllerType;
 use super::ConfigTrait;
+use super::controller::ControllerType;
 
 pub fn create_controller_from_yaml_value_config(
     agent_id: &str,

src/agent/provider/anthropic/utils.rs

@@ -1,8 +1,12 @@
-use anthropic_rs::completion::message::{Content, ContentType, Message, MessageRequest, Role};
+use anthropic::types::{
+    ContentBlock, ImageSource, Message, MessagesRequest, MessagesRequestBuilder, Role,
+};
 
-use crate::conversation::llm::{Author as LLMAuthor, Message as LLMMessage};
+use crate::conversation::llm::{
+    Author as LLMAuthor, Message as LLMMessage, MessageContent as LLMMessageContent,
+};
 
-pub(super) fn create_anthropic_message_request(llm_messages: Vec<LLMMessage>) -> MessageRequest {
+pub(super) fn create_anthropic_message_request(llm_messages: Vec<LLMMessage>) -> MessagesRequest {
     let mut messages = vec![];
 
     for message in llm_messages {
@@ -14,19 +18,26 @@ pub(super) fn create_anthropic_message_request(llm_messages: Vec<LLMMessage>) ->
             }
         };
 
-        let content = vec![Content {
+        let content = match &message.content {
-            content_type: ContentType::Text,
+            LLMMessageContent::Text(text) => vec![ContentBlock::Text { text: text.clone() }],
-            text: message.message_text,
+            LLMMessageContent::Image(image_details) => {
-        }];
+                vec![ContentBlock::Image {
+                    source: ImageSource::Base64 {
+                        media_type: image_details.mime.to_string(),
+                        data: crate::utils::base64::base64_encode(&image_details.data),
+                    },
+                }]
+            }
+        };
 
         let message = Message { role, content };
 
         messages.push(message);
     }
 
-    MessageRequest {
+    MessagesRequestBuilder::default()
-        stream: false,
+        .messages(messages)
-        messages,
+        .stream(false)
-        ..Default::default()
+        .build()
-    }
+        .expect("Failed to build messages request")
 }

src/agent/provider/controller.rs

@@ -1,11 +1,11 @@
 use crate::{agent::AgentPurpose, conversation::llm::Conversation};
 
 use super::{
+    ImageEditParams, ImageGenerationParams, SpeechToTextParams, SpeechToTextResult,
     entity::{
-        ImageGenerationResult, PingResult, TextGenerationParams, TextGenerationResult,
+        ImageEditResult, ImageGenerationResult, ImageSource, PingResult, TextGenerationParams,
-        TextToSpeechParams, TextToSpeechResult,
+        TextGenerationResult, TextToSpeechParams, TextToSpeechResult,
     },
-    ImageGenerationParams, SpeechToTextParams, SpeechToTextResult,
 };
 
 pub trait ControllerTrait {
@@ -42,6 +42,13 @@ pub trait ControllerTrait {
         params: ImageGenerationParams,
     ) -> impl std::future::Future<Output = anyhow::Result<ImageGenerationResult>> + Send;
 
+    fn create_image_edit(
+        &self,
+        prompt: &str,
+        images: Vec<ImageSource>,
+        params: ImageEditParams,
+    ) -> impl std::future::Future<Output = anyhow::Result<ImageEditResult>> + Send;
+
     fn text_to_speech(
         &self,
         text: &str,
@@ -166,6 +173,25 @@ impl ControllerTrait for ControllerType {
         }
     }
 
+    async fn create_image_edit(
+        &self,
+        prompt: &str,
+        images: Vec<ImageSource>,
+        params: ImageEditParams,
+    ) -> anyhow::Result<ImageEditResult> {
+        match &self {
+            ControllerType::OpenAI(controller) => {
+                controller.create_image_edit(prompt, images, params).await
+            }
+            ControllerType::OpenAICompat(controller) => {
+                controller.create_image_edit(prompt, images, params).await
+            }
+            ControllerType::Anthropic(controller) => {
+                controller.create_image_edit(prompt, images, params).await
+            }
+        }
+    }
+
     async fn text_to_speech(
         &self,
         text: &str,

src/agent/provider/entity/agent_provider.rs

@@ -67,9 +67,8 @@ impl AgentProvider {
                 wiki_url: Some("https://en.wikipedia.org/wiki/Anthropic"),
                 sign_up_url: Some("https://console.anthropic.com/"),
                 models_list_url: Some("https://docs.anthropic.com/en/docs/about-claude/models"),
-                supported_purposes: vec![
+                supported_purposes: vec![AgentPurpose::TextGeneration],
-                    AgentPurpose::TextGeneration,
+                text_generation_supports_vision: true,
-                ],
             },
             Self::Groq => AgentProviderInfo {
                 id: Self::Groq.to_static_str(),
@@ -79,10 +78,8 @@ impl AgentProvider {
                 wiki_url: Some("https://en.wikipedia.org/wiki/Groq"),
                 sign_up_url: Some("https://console.groq.com/login"),
                 models_list_url: Some("https://console.groq.com/docs/models"),
-                supported_purposes: vec![
+                supported_purposes: vec![AgentPurpose::TextGeneration, AgentPurpose::SpeechToText],
-                    AgentPurpose::TextGeneration,
+                text_generation_supports_vision: false,
-                    AgentPurpose::SpeechToText,
-                ],
             },
             Self::LocalAI => AgentProviderInfo {
                 id: Self::LocalAI.to_static_str(),
@@ -97,6 +94,7 @@ impl AgentProvider {
                     AgentPurpose::TextToSpeech,
                     AgentPurpose::SpeechToText,
                 ],
+                text_generation_supports_vision: false,
             },
             Self::Mistral => AgentProviderInfo {
                 id: Self::Mistral.to_static_str(),
@@ -106,9 +104,8 @@ impl AgentProvider {
                 wiki_url: Some("https://en.wikipedia.org/wiki/Mistral_AI"),
                 sign_up_url: Some("https://auth.mistral.ai/ui/registration"),
                 models_list_url: Some("https://docs.mistral.ai/getting-started/models/"),
-                supported_purposes: vec![
+                supported_purposes: vec![AgentPurpose::TextGeneration],
-                    AgentPurpose::TextGeneration,
+                text_generation_supports_vision: false,
-                ],
             },
             Self::Ollama => AgentProviderInfo {
                 id: Self::Ollama.to_static_str(),
@@ -118,9 +115,8 @@ impl AgentProvider {
                 wiki_url: None,
                 sign_up_url: None,
                 models_list_url: Some("https://ollama.com/library"),
-                supported_purposes: vec![
+                supported_purposes: vec![AgentPurpose::TextGeneration],
-                    AgentPurpose::TextGeneration,
+                text_generation_supports_vision: false,
-                ],
             },
             Self::OpenAI => AgentProviderInfo {
                 id: Self::OpenAI.to_static_str(),
@@ -136,6 +132,7 @@ impl AgentProvider {
                     AgentPurpose::TextToSpeech,
                     AgentPurpose::SpeechToText,
                 ],
+                text_generation_supports_vision: true,
             },
             Self::OpenAICompat => AgentProviderInfo {
                 id: Self::OpenAICompat.to_static_str(),
@@ -151,6 +148,7 @@ impl AgentProvider {
                     AgentPurpose::TextToSpeech,
                     AgentPurpose::SpeechToText,
                 ],
+                text_generation_supports_vision: false,
             },
             Self::OpenRouter => AgentProviderInfo {
                 id: Self::OpenRouter.to_static_str(),
@@ -160,9 +158,8 @@ impl AgentProvider {
                 wiki_url: None,
                 sign_up_url: Some("https://openrouter.ai/"),
                 models_list_url: Some("https://openrouter.ai/models"),
-                supported_purposes: vec![
+                supported_purposes: vec![AgentPurpose::TextGeneration],
-                    AgentPurpose::TextGeneration,
+                text_generation_supports_vision: false,
-                ],
             },
             Self::TogetherAI => AgentProviderInfo {
                 id: Self::TogetherAI.to_static_str(),
@@ -172,9 +169,8 @@ impl AgentProvider {
                 wiki_url: None,
                 sign_up_url: Some("https://api.together.ai/signup"),
                 models_list_url: Some("https://api.together.xyz/models"),
-                supported_purposes: vec![
+                supported_purposes: vec![AgentPurpose::TextGeneration],
-                    AgentPurpose::TextGeneration,
+                text_generation_supports_vision: false,
-                ],
             },
         }
     }
@@ -195,4 +191,5 @@ pub struct AgentProviderInfo {
     pub sign_up_url: Option<&'static str>,
     pub models_list_url: Option<&'static str>,
     pub supported_purposes: Vec<AgentPurpose>,
+    pub text_generation_supports_vision: bool,
 }

src/agent/provider/entity/image.rs

@@ -1,3 +1,5 @@
+use mxlink::mime;
+
 #[derive(Default)]
 pub struct ImageGenerationParams {
     pub size_override: Option<String>,
@@ -26,6 +28,40 @@ impl ImageGenerationParams {
 
 pub struct ImageGenerationResult {
     pub bytes: Vec<u8>,
-    pub mime_type: mxlink::mime::Mime,
+    pub mime_type: mime::Mime,
     pub revised_prompt: Option<String>,
 }
+
+#[derive(Default)]
+pub struct ImageEditParams {}
+
+pub struct ImageEditResult {
+    pub bytes: Vec<u8>,
+    pub mime_type: mime::Mime,
+}
+
+pub struct ImageSource {
+    pub filename: String,
+    pub bytes: Vec<u8>,
+    pub mime_type: mime::Mime,
+}
+
+impl ImageSource {
+    pub fn new(filename: String, bytes: Vec<u8>, mime_type: mime::Mime) -> Self {
+        Self {
+            filename,
+            bytes,
+            mime_type,
+        }
+    }
+}
+
+impl From<ImageSource> for async_openai::types::ImageInput {
+    fn from(value: ImageSource) -> Self {
+        async_openai::types::ImageInput::from_vec_u8(
+            value.filename,
+            value.bytes,
+            value.mime_type.to_string(),
+        )
+    }
+}

src/agent/provider/entity/mod.rs

@@ -1,12 +1,14 @@
 mod agent_provider;
-mod image_generation;
+mod image;
 mod ping;
 mod speech_to_text;
 mod text_generation;
 mod text_to_speech;
 
 pub use agent_provider::{AgentProvider, AgentProviderInfo};
-pub use image_generation::{ImageGenerationParams, ImageGenerationResult};
+pub use image::{
+    ImageEditParams, ImageEditResult, ImageGenerationParams, ImageGenerationResult, ImageSource,
+};
 pub use ping::PingResult;
 pub use speech_to_text::{SpeechToTextParams, SpeechToTextResult};
 pub use text_generation::{

src/agent/provider/mod.rs

@@ -20,6 +20,7 @@ pub use controller::{ControllerTrait, ControllerType};
 pub use config::ConfigTrait;
 
 pub use entity::{
-    AgentProvider, AgentProviderInfo, ImageGenerationParams, PingResult, SpeechToTextParams,
+    AgentProvider, AgentProviderInfo, ImageEditParams, ImageGenerationParams, ImageSource,
-    SpeechToTextResult, TextGenerationParams, TextGenerationPromptVariables, TextToSpeechParams,
+    PingResult, SpeechToTextParams, SpeechToTextResult, TextGenerationParams,
+    TextGenerationPromptVariables, TextToSpeechParams,
 };

src/agent/provider/openai/config.rs

@@ -1,5 +1,6 @@
 use serde::{Deserialize, Serialize};
 
+use super::OPENAI_IMAGE_MODEL_GPT_IMAGE_1;
 use crate::agent::{default_prompt, provider::ConfigTrait};
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -58,6 +59,9 @@ pub struct TextGenerationConfig {
     #[serde(default)]
     pub max_response_tokens: Option<u32>,
 
+    #[serde(default)]
+    pub max_completion_tokens: Option<u32>,
+
     #[serde(default)]
     pub max_context_tokens: u32,
 }
@@ -69,13 +73,14 @@ impl Default for TextGenerationConfig {
             prompt: Some(default_prompt().to_owned()),
             temperature: super::super::default_temperature(),
             max_response_tokens: Some(16_384),
+            max_completion_tokens: None,
             max_context_tokens: 128_000,
         }
     }
 }
 
 fn default_text_model_id() -> String {
-    "gpt-4o".to_owned()
+    "gpt-4.1".to_owned()
 }
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -145,19 +150,19 @@ pub struct ImageGenerationConfig {
     pub model_id: String,
 
     #[serde(default = "default_image_style")]
-    pub style: async_openai::types::ImageStyle,
+    pub style: Option<async_openai::types::ImageStyle>,
 
     #[serde(default = "default_image_size")]
-    pub size: async_openai::types::ImageSize,
+    pub size: Option<async_openai::types::ImageSize>,
 
     #[serde(default = "default_image_quality")]
-    pub quality: async_openai::types::ImageQuality,
+    pub quality: Option<async_openai::types::ImageQuality>,
 }
 
 impl Default for ImageGenerationConfig {
     fn default() -> Self {
         Self {
-            model_id: "dall-e-3".to_owned(),
+            model_id: OPENAI_IMAGE_MODEL_GPT_IMAGE_1.to_owned(),
             style: default_image_style(),
             size: default_image_size(),
             quality: default_image_quality(),
@@ -177,14 +182,14 @@ impl ImageGenerationConfig {
     }
 }
 
-fn default_image_style() -> async_openai::types::ImageStyle {
+fn default_image_style() -> Option<async_openai::types::ImageStyle> {
-    async_openai::types::ImageStyle::Vivid
+    None
 }
 
-fn default_image_size() -> async_openai::types::ImageSize {
+fn default_image_size() -> Option<async_openai::types::ImageSize> {
-    async_openai::types::ImageSize::S1024x1024
+    None
 }
 
-fn default_image_quality() -> async_openai::types::ImageQuality {
+fn default_image_quality() -> Option<async_openai::types::ImageQuality> {
-    async_openai::types::ImageQuality::Standard
+    None
 }

src/agent/provider/openai/controller.rs

@@ -1,41 +1,45 @@
 use std::ops::Deref;
 
 use async_openai::{
+    Client as OpenAIClient,
     config::OpenAIConfig,
     types::{
-        ChatCompletionRequestMessage, CreateChatCompletionRequestArgs, CreateImageRequestArgs,
+        ChatCompletionRequestMessage, CreateChatCompletionRequestArgs, CreateImageEditRequestArgs,
-        CreateSpeechRequestArgs, CreateTranscriptionRequestArgs,
+        CreateImageRequestArgs, CreateSpeechRequestArgs, CreateTranscriptionRequestArgs,
+        DallE2ImageSize, Image, ImageModel, ImageResponseFormat,
     },
-    Client as OpenAIClient,
 };
 
 use super::super::ControllerTrait;
 use crate::{
-    agent::{
+    agent::provider::{
-        provider::{
+        ImageEditParams, ImageGenerationParams, SpeechToTextParams, SpeechToTextResult,
-            entity::{ImageGenerationResult, PingResult, TextToSpeechParams, TextToSpeechResult},
+        entity::{TextGenerationParams, TextGenerationResult},
-            openai::utils::convert_string_to_enum,
-        },
-        AgentPurpose,
     },
-    strings,
+    conversation::llm::{
+        Author as LLMAuthor, Conversation as LLMConversation, Message as LLMMessage,
+        MessageContent as LLMMessageContent, shorten_messages_list_to_context_size,
+    },
+    utils::base64::base64_decode,
 };
 use crate::{
     agent::{
+        AgentPurpose,
         provider::{
-            entity::{TextGenerationParams, TextGenerationResult},
+            entity::{
-            ImageGenerationParams, SpeechToTextParams, SpeechToTextResult,
+                ImageEditResult, ImageGenerationResult, ImageSource, PingResult,
+                TextToSpeechParams, TextToSpeechResult,
+            },
+            openai::utils::convert_string_to_enum,
         },
-        utils::base64_decode,
-    },
-    conversation::llm::{
-        shorten_messages_list_to_context_size, Author as LLMAuthor,
-        Conversation as LLMConversation, Message as LLMMessage,
     },
+    strings,
 };
 
 use super::config::Config;
 
+use super::OPENAI_IMAGE_MODEL_GPT_IMAGE_1;
+
 #[derive(Debug, Clone)]
 pub struct Controller {
     config: Config,
@@ -62,7 +66,7 @@ impl ControllerTrait for Controller {
 
         let messages = vec![LLMMessage {
             author: LLMAuthor::User,
-            message_text: "Hello!".to_string(),
+            content: LLMMessageContent::Text("Hello!".to_string()),
             timestamp: chrono::Utc::now(),
         }];
 
@@ -99,7 +103,7 @@ impl ControllerTrait for Controller {
         } else {
             Some(LLMMessage {
                 author: LLMAuthor::Prompt,
-                message_text: prompt_text,
+                content: LLMMessageContent::Text(prompt_text),
                 timestamp: chrono::Utc::now(),
             })
         };
@@ -144,6 +148,10 @@ impl ControllerTrait for Controller {
             request_builder.max_tokens(max_response_tokens);
         }
 
+        if let Some(max_completion_tokens) = text_generation_config.max_completion_tokens {
+            request_builder.max_completion_tokens(max_completion_tokens);
+        }
+
         let request = request_builder.build()?;
 
         if let Ok(request_as_json) = serde_json::to_string(&request) {
@@ -201,12 +209,11 @@ impl ControllerTrait for Controller {
 
         let request = CreateTranscriptionRequestArgs::default()
             .model(&speech_to_text_config.model_id)
-            .file(async_openai::types::AudioInput {
+            .file(async_openai::types::AudioInput::from_vec_u8(
-                source: async_openai::types::InputSource::VecU8 {
+                filename,
-                    filename,
+                media,
-                    vec: media,
+                mime_type.to_string(),
-                },
+            ))
-            })
             .language(language.clone())
             .build()?;
 
@@ -261,12 +268,15 @@ impl ControllerTrait for Controller {
         let quality = if params.cheaper_quality_switching_allowed {
             // Switch to a cheaper quality
             match &image_generation_config.quality {
-                async_openai::types::ImageQuality::Standard => {
+                Some(quality) => match quality {
-                    async_openai::types::ImageQuality::Standard
+                    async_openai::types::ImageQuality::Standard => {
-                }
+                        Some(async_openai::types::ImageQuality::Standard)
-                async_openai::types::ImageQuality::HD => {
+                    }
-                    async_openai::types::ImageQuality::Standard
+                    async_openai::types::ImageQuality::HD => {
-                }
+                        Some(async_openai::types::ImageQuality::Standard)
+                    }
+                },
+                None => None,
             }
         } else {
             image_generation_config.quality.clone()
@@ -274,20 +284,41 @@ impl ControllerTrait for Controller {
 
         let size = params
             .size_override
-            .map(|s| {
+            .map(|s| convert_string_to_enum::<async_openai::types::ImageSize>(&s).unwrap())
-                convert_string_to_enum::<async_openai::types::ImageSize>(&s)
+            .or(image_generation_config.size);
-                    .unwrap_or(image_generation_config.size)
-            })
-            .unwrap_or(image_generation_config.size);
 
-        let request = CreateImageRequestArgs::default()
+        let response_format = match model.clone() {
-            .model(model)
+            ImageModel::DallE2 => Some(ImageResponseFormat::B64Json),
-            .prompt(prompt.to_owned())
+            ImageModel::DallE3 => Some(ImageResponseFormat::B64Json),
-            .response_format(async_openai::types::ImageResponseFormat::B64Json)
+            ImageModel::Other(model_str) => match model_str.as_str() {
-            .size(size)
+                // gpt-image-1 only outputs base64 and we don't need to specify the response format.
-            .style(image_generation_config.style.clone())
+                // In fact, specifying the response format results in an error.
-            .quality(quality)
+                OPENAI_IMAGE_MODEL_GPT_IMAGE_1 => None,
-            .build()?;
+                _ => Some(ImageResponseFormat::B64Json),
+            },
+        };
+
+        let mut request_builder = CreateImageRequestArgs::default();
+
+        request_builder.model(model).prompt(prompt.to_owned());
+
+        if let Some(response_format) = response_format {
+            request_builder.response_format(response_format);
+        }
+
+        if let Some(style) = &image_generation_config.style {
+            request_builder.style(style.clone());
+        }
+
+        if let Some(quality) = quality {
+            request_builder.quality(quality.clone());
+        }
+
+        if let Some(size) = size {
+            request_builder.size(size);
+        }
+
+        let request = request_builder.build()?;
 
         tracing::trace!(
             ?prompt,
@@ -325,6 +356,104 @@ impl ControllerTrait for Controller {
         ))
     }
 
+    async fn create_image_edit(
+        &self,
+        prompt: &str,
+        images: Vec<ImageSource>,
+        _params: ImageEditParams,
+    ) -> anyhow::Result<ImageEditResult> {
+        let Some(image_generation_config) = &self.config.image_generation else {
+            return Err(anyhow::anyhow!(
+                strings::agent::no_configuration_for_purpose_so_cannot_be_used(
+                    &AgentPurpose::ImageGeneration
+                ),
+            ));
+        };
+
+        if images.is_empty() {
+            return Err(anyhow::anyhow!("No image sources provided"));
+        }
+
+        let mut image_inputs = Vec::new();
+        for image in images {
+            image_inputs.push(image.into());
+        }
+
+        let dalle2_size = match image_generation_config.size {
+            Some(async_openai::types::ImageSize::S256x256) => Some(DallE2ImageSize::S256x256),
+            Some(async_openai::types::ImageSize::S512x512) => Some(DallE2ImageSize::S512x512),
+            Some(async_openai::types::ImageSize::S1024x1024) => Some(DallE2ImageSize::S1024x1024),
+            _ => None,
+        };
+
+        let model = image_generation_config
+            .model_id_as_openai_image_model()
+            .map_err(|err| anyhow::anyhow!(err))?;
+
+        let response_format = match model.clone() {
+            async_openai::types::ImageModel::DallE2 => {
+                Some(async_openai::types::ImageResponseFormat::B64Json)
+            }
+            async_openai::types::ImageModel::DallE3 => {
+                Some(async_openai::types::ImageResponseFormat::B64Json)
+            }
+            async_openai::types::ImageModel::Other(model_str) => match model_str.as_str() {
+                OPENAI_IMAGE_MODEL_GPT_IMAGE_1 => None,
+                _ => Some(async_openai::types::ImageResponseFormat::B64Json),
+            },
+        };
+
+        let mut request_builder = CreateImageEditRequestArgs::default();
+
+        request_builder
+            .image(image_inputs)
+            .prompt(prompt.to_owned())
+            .model(model);
+
+        if let Some(size) = dalle2_size {
+            request_builder.size(size);
+        }
+
+        if let Some(response_format) = response_format {
+            request_builder.response_format(response_format);
+        }
+
+        let request = request_builder
+            .build()
+            .map_err(|e| anyhow::anyhow!("Failed to build CreateImageEditRequest: {}", e))?;
+
+        tracing::trace!(
+            model = format!("{:?}", request.model),
+            size = format!("{:?}", request.size),
+            response_format = format!("{:?}", request.response_format),
+            "Sending OpenAI image edit API request"
+        );
+
+        let response = self.client.images().create_edit(request).await?;
+
+        if let Some(image_data) = response.data.into_iter().next() {
+            match image_data.deref() {
+                Image::B64Json { b64_json, .. } => {
+                    let bytes = base64_decode(b64_json)?;
+                    return Ok(ImageEditResult {
+                        bytes,
+                        mime_type: mxlink::mime::IMAGE_PNG,
+                    });
+                }
+                Image::Url { url, .. } => {
+                    tracing::warn!(?url, "Received URL instead of B64Json for image edit");
+                    return Err(anyhow::anyhow!(
+                        "Unexpected image type (URL) when B64Json was requested"
+                    ));
+                }
+            }
+        }
+
+        Err(anyhow::anyhow!(
+            "The OpenAI image edit API returned no images"
+        ))
+    }
+
     async fn text_to_speech(
         &self,
         input: &str,

src/agent/provider/openai/mod.rs

@@ -13,8 +13,10 @@ pub(super) use config::TextToSpeechConfig;
 
 use super::super::AgentInstantiationError;
 use super::super::AgentInstantiationResult;
-use super::controller::ControllerType;
 use super::ConfigTrait;
+use super::controller::ControllerType;
+
+pub const OPENAI_IMAGE_MODEL_GPT_IMAGE_1: &str = "gpt-image-1";
 
 pub fn create_controller_from_yaml_value_config(
     agent_id: &str,

src/agent/provider/openai/utils.rs

@@ -1,9 +1,14 @@
 use async_openai::types::{
     ChatCompletionRequestAssistantMessageArgs, ChatCompletionRequestMessage,
-    ChatCompletionRequestSystemMessageArgs, ChatCompletionRequestUserMessageArgs,
+    ChatCompletionRequestMessageContentPartImage, ChatCompletionRequestSystemMessageArgs,
+    ChatCompletionRequestUserMessageArgs, ChatCompletionRequestUserMessageContent,
+    ChatCompletionRequestUserMessageContentPart, ImageUrlArgs,
 };
 
-use crate::conversation::llm::{Author as LLMAuthor, Message as LLMMessage};
+use crate::conversation::llm::{
+    Author as LLMAuthor, Message as LLMMessage, MessageContent as LLMMessageContent,
+};
+use crate::utils::base64::base64_encode;
 
 pub fn convert_llm_messages_to_openai_messages(
     conversation_messages: Vec<LLMMessage>,
@@ -12,29 +17,71 @@ pub fn convert_llm_messages_to_openai_messages(
         Vec::with_capacity(conversation_messages.len());
 
     for message in conversation_messages {
-        openai_conversation_messages.push(convert_llm_message_to_openai_message(message));
+        let openai_message = convert_llm_message_to_openai_message(message);
+        if let Some(openai_message) = openai_message {
+            openai_conversation_messages.push(openai_message);
+        }
     }
 
     openai_conversation_messages
 }
 
-fn convert_llm_message_to_openai_message(llm_message: LLMMessage) -> ChatCompletionRequestMessage {
+fn convert_llm_message_to_openai_message(
-    match llm_message.author {
+    llm_message: LLMMessage,
-        LLMAuthor::Prompt => ChatCompletionRequestSystemMessageArgs::default()
+) -> Option<ChatCompletionRequestMessage> {
-            .content(llm_message.message_text)
+    match &llm_message.content {
-            .build()
+        LLMMessageContent::Text(text) => Some(match llm_message.author {
-            .expect("Failed building OpenAI system message")
+            LLMAuthor::Prompt => ChatCompletionRequestSystemMessageArgs::default()
-            .into(),
+                .content(text.clone())
-        LLMAuthor::Assistant => ChatCompletionRequestAssistantMessageArgs::default()
+                .build()
-            .content(llm_message.message_text)
+                .expect("Failed building OpenAI system message")
-            .build()
+                .into(),
-            .expect("Failed building OpenAI assistant message")
+            LLMAuthor::Assistant => ChatCompletionRequestAssistantMessageArgs::default()
-            .into(),
+                .content(text.clone())
-        LLMAuthor::User => ChatCompletionRequestUserMessageArgs::default()
+                .build()
-            .content(llm_message.message_text)
+                .expect("Failed building OpenAI assistant message")
-            .build()
+                .into(),
-            .expect("Failed building OpenAI user message")
+            LLMAuthor::User => ChatCompletionRequestUserMessageArgs::default()
-            .into(),
+                .content(text.clone())
+                .build()
+                .expect("Failed building OpenAI user message")
+                .into(),
+        }),
+        LLMMessageContent::Image(image_details) => {
+            let image_url = format!(
+                "data:{};base64,{}",
+                image_details.mime,
+                base64_encode(&image_details.data)
+            );
+
+            let part = ChatCompletionRequestUserMessageContentPart::ImageUrl(
+                ChatCompletionRequestMessageContentPartImage {
+                    image_url: ImageUrlArgs::default()
+                        .url(image_url)
+                        .build()
+                        .expect("Failed building OpenAI image url"),
+                },
+            );
+
+            let message_content = ChatCompletionRequestUserMessageContent::Array(vec![part]);
+
+            match llm_message.author {
+                LLMAuthor::User => Some(
+                    ChatCompletionRequestUserMessageArgs::default()
+                        .content(message_content)
+                        .build()
+                        .expect("Failed building OpenAI user message")
+                        .into(),
+                ),
+                _ => {
+                    tracing::warn!(
+                        "OpenAI API does not support image content for messages authored by {:?}. This message part will be skipped.",
+                        llm_message.author
+                    );
+                    None
+                }
+            }
+        }
     }
 }
 

src/agent/provider/openai_compat/config.rs

@@ -93,6 +93,7 @@ impl TryInto<OpenAITextGenerationConfig> for TextGenerationConfig {
             prompt: self.prompt,
             temperature: self.temperature,
             max_response_tokens: self.max_response_tokens,
+            max_completion_tokens: None,
             max_context_tokens: self.max_context_tokens,
         })
     }
@@ -223,21 +224,27 @@ impl TryInto<OpenAIImageGenerationConfig> for ImageGenerationConfig {
 
     fn try_into(self) -> Result<OpenAIImageGenerationConfig, Self::Error> {
         let size = if let Some(size) = &self.size {
-            convert_string_to_enum::<async_openai::types::ImageSize>(size)?
+            Some(convert_string_to_enum::<async_openai::types::ImageSize>(
+                size,
+            )?)
         } else {
-            async_openai::types::ImageSize::S1024x1024
+            None
         };
 
         let style = if let Some(style) = &self.style {
-            convert_string_to_enum::<async_openai::types::ImageStyle>(style)?
+            Some(convert_string_to_enum::<async_openai::types::ImageStyle>(
+                style,
+            )?)
         } else {
-            async_openai::types::ImageStyle::Vivid
+            None
         };
 
         let quality = if let Some(quality) = &self.quality {
-            convert_string_to_enum::<async_openai::types::ImageQuality>(quality)?
+            Some(convert_string_to_enum::<async_openai::types::ImageQuality>(
+                quality,
+            )?)
         } else {
-            async_openai::types::ImageQuality::Standard
+            None
         };
 
         Ok(OpenAIImageGenerationConfig {

src/agent/provider/openai_compat/controller.rs

@@ -4,23 +4,25 @@ use etke_openai_api_rust::images::{ImagesApi, ImagesBody};
 use etke_openai_api_rust::{Auth, Message, OpenAI};
 
 use super::super::ControllerTrait;
-use crate::agent::utils::base64_decode;
+use crate::utils::base64::base64_decode;
 use crate::{
     agent::provider::{
+        ImageEditParams, ImageGenerationParams, ImageSource, SpeechToTextParams,
+        SpeechToTextResult,
         entity::{TextGenerationParams, TextGenerationResult},
-        ImageGenerationParams, SpeechToTextParams, SpeechToTextResult,
     },
     conversation::llm::{
-        shorten_messages_list_to_context_size, Author as LLMAuthor,
+        Author as LLMAuthor, Conversation as LLMConversation, Message as LLMMessage,
-        Conversation as LLMConversation, Message as LLMMessage,
+        MessageContent as LLMMessageContent, shorten_messages_list_to_context_size,
     },
 };
 use crate::{
     agent::{
-        provider::entity::{
-            ImageGenerationResult, PingResult, TextToSpeechParams, TextToSpeechResult,
-        },
         AgentPurpose,
+        provider::entity::{
+            ImageEditResult, ImageGenerationResult, PingResult, TextToSpeechParams,
+            TextToSpeechResult,
+        },
     },
     strings,
 };
@@ -60,7 +62,7 @@ impl ControllerTrait for Controller {
 
         let messages = vec![LLMMessage {
             author: LLMAuthor::User,
-            message_text: "Hello!".to_string(),
+            content: LLMMessageContent::Text("Hello!".to_string()),
             timestamp: chrono::Utc::now(),
         }];
 
@@ -97,7 +99,7 @@ impl ControllerTrait for Controller {
         } else {
             Some(LLMMessage {
                 author: LLMAuthor::Prompt,
-                message_text: prompt_text,
+                content: LLMMessageContent::Text(prompt_text),
                 timestamp: chrono::Utc::now(),
             })
         };
@@ -366,6 +368,17 @@ impl ControllerTrait for Controller {
         ))
     }
 
+    async fn create_image_edit(
+        &self,
+        _prompt: &str,
+        _images: Vec<ImageSource>,
+        _params: ImageEditParams,
+    ) -> anyhow::Result<ImageEditResult> {
+        Err(anyhow::anyhow!(
+            "The OpenAI image edit API is not supported by the OpenAI-compat provider"
+        ))
+    }
+
     async fn text_to_speech(
         &self,
         input: &str,

src/agent/provider/openai_compat/mod.rs

@@ -21,8 +21,8 @@ pub use controller::Controller;
 
 use super::super::AgentInstantiationError;
 use super::super::AgentInstantiationResult;
-use super::controller::ControllerType;
 use super::ConfigTrait;
+use super::controller::ControllerType;
 
 pub fn create_controller_from_yaml_value_config(
     agent_id: &str,

src/agent/provider/openai_compat/utils.rs

@@ -2,7 +2,9 @@ use etke_openai_api_rust::{Message, Role};
 
 use crate::agent::provider::openai::Config as OpenAIConfig;
 
-use crate::conversation::llm::{Author as LLMAuthor, Message as LLMMessage};
+use crate::conversation::llm::{
+    Author as LLMAuthor, Message as LLMMessage, MessageContent as LLMMessageContent,
+};
 
 pub fn convert_llm_messages_to_openai_messages(
     conversation_messages: Vec<LLMMessage>,
@@ -11,22 +13,33 @@ pub fn convert_llm_messages_to_openai_messages(
         Vec::with_capacity(conversation_messages.len());
 
     for message in conversation_messages {
-        openai_conversation_messages.push(convert_llm_message_to_openai_message(message));
+        let openai_message = convert_llm_message_to_openai_message(message);
+        if let Some(openai_message) = openai_message {
+            openai_conversation_messages.push(openai_message);
+        }
     }
 
     openai_conversation_messages
 }
 
-fn convert_llm_message_to_openai_message(llm_message: LLMMessage) -> Message {
+fn convert_llm_message_to_openai_message(llm_message: LLMMessage) -> Option<Message> {
     let role = match llm_message.author {
         LLMAuthor::Prompt => Role::System,
         LLMAuthor::Assistant => Role::Assistant,
         LLMAuthor::User => Role::User,
     };
 
-    Message {
+    match &llm_message.content {
-        role,
+        LLMMessageContent::Text(text) => Some(Message {
-        content: llm_message.message_text,
+            role,
+            content: text.clone(),
+        }),
+        LLMMessageContent::Image(_image_details) => {
+            tracing::warn!(
+                "The OpenAI-compat provider's library does not support image content. This image message will be skipped."
+            );
+            None
+        }
     }
 }
 

src/agent/utils.rs

@@ -1,5 +1,3 @@
-use base64::{engine::general_purpose::STANDARD, Engine as _};
-
 use crate::{
     agent::{
         AgentInstance, AgentPurpose, ControllerTrait, Manager as AgentManager, PublicIdentifier,
@@ -140,7 +138,3 @@ async fn get_global_agent_id_for_purpose(
         .handler
         .get_by_purpose_with_catch_all_fallback(purpose)
 }
-
-pub(crate) fn base64_decode(base64_string: &str) -> Result<Vec<u8>, base64::DecodeError> {
-    STANDARD.decode(base64_string)
-}

src/bot/implementation.rs

@@ -1,11 +1,11 @@
 use std::sync::Arc;
 use std::{future::Future, pin::Pin};
 
+use mxlink::matrix_sdk::Room;
 use mxlink::matrix_sdk::media::{MediaFormat, MediaRequestParameters};
 use mxlink::matrix_sdk::ruma::{
-    events::room::MediaSource, MilliSecondsSinceUnixEpoch, OwnedUserId,
+    MilliSecondsSinceUnixEpoch, OwnedUserId, events::room::MediaSource,
 };
-use mxlink::matrix_sdk::Room;
 
 use mxlink::{
     InitConfig, LoginConfig, LoginCredentials, LoginEncryption, MatrixLink, PersistenceConfig,
@@ -140,6 +140,10 @@ impl Bot {
         &self.inner.config.command_prefix
     }
 
+    pub(crate) fn post_join_self_introduction_enabled(&self) -> bool {
+        self.inner.config.room.post_join_self_introduction_enabled
+    }
+
     pub(crate) fn homeserver_name(&self) -> &str {
         &self.inner.config.homeserver.server_name
     }

src/bot/load_config.rs

@@ -5,7 +5,7 @@ use anyhow::anyhow;
 
 use crate::agent::AgentPurpose;
 
-pub use crate::entity::cfg::{defaults as cfg_defaults, env as cfg_env, Config};
+pub use crate::entity::cfg::{Config, defaults as cfg_defaults, env as cfg_env};
 
 pub fn load() -> anyhow::Result<Config> {
     let config_file_path = env::var(cfg_env::BAIBOT_CONFIG_FILE_PATH)
@@ -35,6 +35,9 @@ pub fn load() -> anyhow::Result<Config> {
             }
             cfg_env::BAIBOT_USER_NAME => config.user.name = value,
             cfg_env::BAIBOT_COMMAND_PREFIX => config.command_prefix = value,
+            cfg_env::BAIBOT_ROOM_POST_JOIN_SELF_INTRODUCTION_ENABLED => {
+                config.room.post_join_self_introduction_enabled = value.parse::<bool>()?;
+            }
             cfg_env::BAIBOT_LOGGING => {
                 config.logging = value;
             }

src/bot/messaging.rs

@@ -1,9 +1,9 @@
 use mxlink::matrix_sdk::{
-    ruma::{
-        api::client::receipt::create_receipt::v3::ReceiptType,
-        events::room::message::OriginalSyncRoomMessageEvent, OwnedEventId,
-    },
     Room,
+    ruma::{
+        OwnedEventId, api::client::receipt::create_receipt::v3::ReceiptType,
+        events::room::message::OriginalSyncRoomMessageEvent,
+    },
 };
 
 use mxlink::{CallbackError, MessageResponseType};

src/bot/reacting.rs

@@ -1,12 +1,12 @@
 use mxlink::matrix_sdk::{
-    ruma::{
-        events::{
-            room::message::Relation, AnySyncMessageLikeEvent, AnySyncTimelineEvent,
-            SyncMessageLikeEvent,
-        },
-        OwnedEventId, OwnedUserId,
-    },
     Room,
+    ruma::{
+        OwnedEventId, OwnedUserId,
+        events::{
+            AnySyncMessageLikeEvent, AnySyncTimelineEvent, SyncMessageLikeEvent,
+            room::message::Relation,
+        },
+    },
 };
 
 use mxlink::CallbackError;

src/bot/rooms.rs

@@ -1,9 +1,9 @@
 use mxlink::{
-    matrix_sdk::{
-        ruma::events::{room::member::StrippedRoomMemberEvent, AnySyncTimelineEvent},
-        Room,
-    },
     InvitationDecision,
+    matrix_sdk::{
+        Room,
+        ruma::events::{AnySyncTimelineEvent, room::member::StrippedRoomMemberEvent},
+    },
 };
 
 use mxlink::CallbackError;

src/controller/access/dispatching.rs

@@ -1,6 +1,6 @@
 use mxlink::MessageResponseType;
 
-use crate::{entity::MessageContext, strings, Bot};
+use crate::{Bot, entity::MessageContext, strings};
 
 use super::AccessControllerType;
 

src/controller/access/help.rs

@@ -1,6 +1,6 @@
 use mxlink::MessageResponseType;
 
-use crate::{entity::MessageContext, strings, Bot};
+use crate::{Bot, entity::MessageContext, strings};
 
 pub async fn handle(bot: &Bot, message_context: &MessageContext) -> anyhow::Result<()> {
     let mut message = String::new();

src/controller/access/mod.rs

@@ -4,5 +4,5 @@ pub mod help;
 mod room_local_agent_managers;
 mod users;
 
-pub use determination::{determine_controller, AccessControllerType};
+pub use determination::{AccessControllerType, determine_controller};
 pub use dispatching::dispatch_controller;

src/controller/access/room_local_agent_managers.rs

@@ -1,6 +1,6 @@
 use mxlink::MessageResponseType;
 
-use crate::{entity::MessageContext, strings, Bot};
+use crate::{Bot, entity::MessageContext, strings};
 
 pub async fn handle_get(bot: &Bot, message_context: &MessageContext) -> anyhow::Result<()> {
     let message = match &message_context

src/controller/access/users.rs

@@ -1,6 +1,6 @@
 use mxlink::MessageResponseType;
 
-use crate::{entity::MessageContext, strings, Bot};
+use crate::{Bot, entity::MessageContext, strings};
 
 pub async fn handle_get(bot: &Bot, message_context: &MessageContext) -> anyhow::Result<()> {
     let message = match &message_context.global_config().access.user_patterns {

src/controller/agent/create/mod.rs

@@ -3,15 +3,15 @@ mod tests;
 
 use mxlink::MessageResponseType;
 
-use crate::agent::provider::{ControllerTrait, PingResult};
 use crate::agent::PublicIdentifier;
-use crate::agent::{create_from_provider_and_yaml_value_config, AgentDefinition};
+use crate::agent::provider::{ControllerTrait, PingResult};
+use crate::agent::{AgentDefinition, create_from_provider_and_yaml_value_config};
 use crate::agent::{AgentInstance, AgentProvider};
 use crate::controller::utils::get_text_body_or_complain;
 use crate::entity::globalconfig::GlobalConfigurationManager;
 use crate::entity::roomconfig::RoomConfigurationManager;
 use crate::strings;
-use crate::{entity::MessageContext, Bot};
+use crate::{Bot, entity::MessageContext};
 
 struct ParsedAgentConfig {
     agent: AgentInstance,

src/controller/agent/delete/mod.rs

@@ -1,9 +1,9 @@
 use mxlink::MessageResponseType;
 
 use crate::entity::{
-    globalconfig::GlobalConfigurationManager, roomconfig::RoomConfigurationManager, MessageContext,
+    MessageContext, globalconfig::GlobalConfigurationManager, roomconfig::RoomConfigurationManager,
 };
-use crate::{agent::PublicIdentifier, strings, Bot};
+use crate::{Bot, agent::PublicIdentifier, strings};
 
 pub async fn handle(
     bot: &Bot,

src/controller/agent/details/mod.rs

@@ -1,6 +1,6 @@
 use mxlink::MessageResponseType;
 
-use crate::{agent::PublicIdentifier, entity::MessageContext, strings, Bot};
+use crate::{Bot, agent::PublicIdentifier, entity::MessageContext, strings};
 
 pub async fn handle(
     bot: &Bot,

src/controller/agent/help/mod.rs

@@ -1,6 +1,6 @@
 use mxlink::MessageResponseType;
 
-use crate::{entity::MessageContext, strings, Bot};
+use crate::{Bot, entity::MessageContext, strings};
 
 pub async fn handle(bot: &Bot, message_context: &MessageContext) -> anyhow::Result<()> {
     // Anyone can access this help command, because certain subcommands ("list")

src/controller/agent/list/mod.rs

@@ -2,7 +2,7 @@ use mxlink::MessageResponseType;
 
 use crate::agent::AgentPurpose;
 use crate::strings;
-use crate::{entity::MessageContext, Bot};
+use crate::{Bot, entity::MessageContext};
 
 pub async fn handle(bot: &Bot, message_context: &MessageContext) -> anyhow::Result<()> {
     let agents = bot

src/controller/agent/mod.rs

@@ -1,4 +1,4 @@
-use crate::{entity::MessageContext, Bot};
+use crate::{Bot, entity::MessageContext};
 
 pub mod create;
 pub mod delete;
@@ -7,7 +7,7 @@ pub mod determination;
 pub mod help;
 pub mod list;
 
-pub use determination::{determine_controller, AgentControllerType};
+pub use determination::{AgentControllerType, determine_controller};
 
 pub async fn dispatch_controller(
     handler: &AgentControllerType,

src/controller/cfg/common/generic_setting.rs

@@ -1,6 +1,6 @@
 use mxlink::MessageResponseType;
 
-use crate::{entity::MessageContext, strings, Bot};
+use crate::{Bot, entity::MessageContext, strings};
 
 pub async fn handle_get<T>(
     bot: &Bot,

src/controller/cfg/controller_type.rs

@@ -1,7 +1,8 @@
 use crate::{
     agent::{AgentPurpose, PublicIdentifier},
     entity::roomconfig::{
-        SpeechToTextFlowType, TextGenerationAutoUsage, TextGenerationPrefixRequirementType,
+        SpeechToTextFlowType, SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages,
+        TextGenerationAutoUsage, TextGenerationPrefixRequirementType,
         TextToSpeechBotMessagesFlowType, TextToSpeechUserMessagesFlowType,
     },
 };
@@ -54,6 +55,11 @@ pub enum ConfigSpeechToTextSettingRelatedControllerType {
     GetFlowType,
     SetFlowType(Option<SpeechToTextFlowType>),
 
+    GetMsgTypeForNonThreadedOnlyTranscribedMessages,
+    SetMsgTypeForNonThreadedOnlyTranscribedMessages(
+        Option<SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages>,
+    ),
+
     GetLanguage,
     SetLanguage(Option<String>),
 }

src/controller/cfg/determination/speech_to_text/mod.rs

@@ -1,7 +1,13 @@
 #[cfg(test)]
 mod tests;
 
-use crate::{controller::ControllerType, entity::roomconfig::SpeechToTextFlowType, strings};
+use crate::{
+    controller::ControllerType,
+    entity::roomconfig::{
+        SpeechToTextFlowType, SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages,
+    },
+    strings,
+};
 
 use super::super::controller_type::ConfigSpeechToTextSettingRelatedControllerType;
 
@@ -48,6 +54,53 @@ pub(super) fn determine(
         ));
     }
 
+    // msg_type_for_non_threaded_only_transcribed_messages
+
+    if let Some(remaining_text) =
+        text.strip_prefix("msg-type-for-non-threaded-only-transcribed-messages")
+    {
+        let remaining_text = remaining_text.trim();
+
+        if !remaining_text.is_empty() {
+            return Err(ControllerType::Error(
+                strings::cfg::configuration_getter_used_with_extra_text(
+                    "msg-type-for-non-threaded-only-transcribed-messages",
+                    remaining_text,
+                )
+                .to_owned(),
+            ));
+        }
+
+        return Ok(ConfigSpeechToTextSettingRelatedControllerType::GetMsgTypeForNonThreadedOnlyTranscribedMessages);
+    }
+
+    if let Some(value_string) =
+        text.strip_prefix("set-msg-type-for-non-threaded-only-transcribed-messages")
+    {
+        let value_string = value_string.trim().to_owned();
+
+        let value_choice = if value_string.is_empty() {
+            None
+        } else {
+            let value_choice =
+                SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages::from_str(
+                    &value_string.to_lowercase(),
+                );
+
+            if value_choice.is_none() {
+                return Err(ControllerType::Error(
+                    strings::cfg::configuration_value_unrecognized(&value_string).to_owned(),
+                ));
+            }
+
+            value_choice
+        };
+
+        return Ok(ConfigSpeechToTextSettingRelatedControllerType::SetMsgTypeForNonThreadedOnlyTranscribedMessages(
+            value_choice,
+        ));
+    }
+
     // Language
 
     if let Some(remaining_text) = text.strip_prefix("language") {

src/controller/cfg/dispatching/mod.rs

@@ -1,5 +1,5 @@
 use crate::strings;
-use crate::{entity::MessageContext, Bot};
+use crate::{Bot, entity::MessageContext};
 use mxlink::MessageResponseType;
 
 use super::controller_type::{

src/controller/cfg/dispatching/speech_to_text.rs

@@ -1,5 +1,8 @@
-use crate::entity::roomconfig::{RoomSettings, SpeechToTextFlowType};
+use crate::entity::roomconfig::{
-use crate::{entity::MessageContext, Bot};
+    RoomSettings, SpeechToTextFlowType,
+    SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages,
+};
+use crate::{Bot, entity::MessageContext};
 
 use super::super::controller_type::{
     ConfigSpeechToTextSettingRelatedControllerType, SettingsStorageSource,
@@ -52,6 +55,39 @@ pub(super) async fn dispatch(
             }
         }
 
+        ConfigSpeechToTextSettingRelatedControllerType::GetMsgTypeForNonThreadedOnlyTranscribedMessages => {
+            let value = &room_settings.speech_to_text.msg_type_for_non_threaded_only_transcribed_messages;
+            setting_get::<SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages>(bot, message_context, value).await
+        }
+        ConfigSpeechToTextSettingRelatedControllerType::SetMsgTypeForNonThreadedOnlyTranscribedMessages(value) => {
+            let value = value.to_owned();
+
+            let setter_callback = Box::new(move |room_settings: &mut RoomSettings| {
+                room_settings.speech_to_text.msg_type_for_non_threaded_only_transcribed_messages = value;
+            });
+
+            match config_type {
+                SettingsStorageSource::Room => {
+                    room_setting_set::<SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages>(
+                        bot,
+                        message_context,
+                        &value,
+                        setter_callback,
+                    )
+                    .await
+                }
+                SettingsStorageSource::Global => {
+                    global_setting_set::<SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages>(
+                        bot,
+                        message_context,
+                        &value,
+                        setter_callback,
+                    )
+                    .await
+                }
+            }
+        }
+
         ConfigSpeechToTextSettingRelatedControllerType::GetLanguage => {
             let value = &room_settings.speech_to_text.language;
             setting_get::<String>(bot, message_context, value).await

src/controller/cfg/dispatching/text_generation.rs

@@ -1,7 +1,7 @@
 use crate::entity::roomconfig::{
     RoomSettings, TextGenerationAutoUsage, TextGenerationPrefixRequirementType,
 };
-use crate::{entity::MessageContext, Bot};
+use crate::{Bot, entity::MessageContext};
 
 use super::super::controller_type::{
     ConfigTextGenerationSettingRelatedControllerType, SettingsStorageSource,

src/controller/cfg/dispatching/text_to_speech.rs

@@ -1,7 +1,7 @@
 use crate::entity::roomconfig::{
     RoomSettings, TextToSpeechBotMessagesFlowType, TextToSpeechUserMessagesFlowType,
 };
-use crate::{entity::MessageContext, Bot};
+use crate::{Bot, entity::MessageContext};
 
 use super::super::controller_type::{
     ConfigTextToSpeechSettingRelatedControllerType, SettingsStorageSource,

src/controller/cfg/global_config/generic_setting.rs

@@ -1,7 +1,7 @@
 use mxlink::MessageResponseType;
 
-use crate::entity::{roomconfig::RoomSettings, MessageContext};
+use crate::entity::{MessageContext, roomconfig::RoomSettings};
-use crate::{strings, Bot};
+use crate::{Bot, strings};
 
 pub async fn handle_set<T>(
     bot: &Bot,

src/controller/cfg/global_config/handler.rs

@@ -1,9 +1,10 @@
 use mxlink::MessageResponseType;
 
 use crate::{
+    Bot,
     agent::{AgentPurpose, PublicIdentifier},
-    entity::{globalconfig::GlobalConfigurationManager, MessageContext},
+    entity::{MessageContext, globalconfig::GlobalConfigurationManager},
-    strings, Bot,
+    strings,
 };
 
 pub async fn handle_get(

src/controller/cfg/help.rs

@@ -1,14 +1,16 @@
 use mxlink::MessageResponseType;
 
 use crate::{
+    Bot,
     entity::{
+        MessageContext,
         roomconfig::{
-            SpeechToTextFlowType, TextGenerationAutoUsage, TextGenerationPrefixRequirementType,
+            SpeechToTextFlowType, SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages,
+            TextGenerationAutoUsage, TextGenerationPrefixRequirementType,
             TextToSpeechBotMessagesFlowType, TextToSpeechUserMessagesFlowType,
         },
-        MessageContext,
     },
-    strings, Bot,
+    strings,
 };
 
 pub async fn handle(bot: &Bot, message_context: &MessageContext) -> anyhow::Result<()> {
@@ -346,6 +348,46 @@ fn build_section_speech_to_text(command_prefix: &str) -> String {
     ));
     message.push_str("\n\n");
 
+    // Msg Type For Non Threaded Only Transcribed Messages
+
+    message.push_str(&format!(
+        "#### {}",
+        strings::help::cfg::speech_to_text_msg_type_for_non_threaded_only_transcribed_messages_heading()
+    ));
+    message.push_str("\n\n");
+    message.push_str(strings::help::cfg::speech_to_text_msg_type_for_non_threaded_only_transcribed_messages_intro());
+    message.push('\n');
+    message.push_str(
+        &strings::help::cfg::the_following_configuration_values_are_recognized(
+            SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages::choices(),
+        ),
+    );
+    message.push_str("\n\n");
+    message.push_str(&format!(
+        "- {}",
+        &strings::help::cfg::current_setting_show(
+            command_prefix,
+            "speech-to-text msg-type-for-non-threaded-only-transcribed-messages"
+        )
+    ));
+    message.push('\n');
+    message.push_str(&format!(
+        "- {}",
+        &strings::help::cfg::current_setting_set(
+            command_prefix,
+            "speech-to-text set-msg-type-for-non-threaded-only-transcribed-messages VALUE"
+        )
+    ));
+    message.push('\n');
+    message.push_str(&format!(
+        "- {}",
+        &strings::help::cfg::current_setting_unset(
+            command_prefix,
+            "speech-to-text set-msg-type-for-non-threaded-only-transcribed-messages"
+        )
+    ));
+    message.push_str("\n\n");
+
     // Language
 
     message.push_str(&format!(

src/controller/cfg/room_config/generic_setting.rs

@@ -1,7 +1,7 @@
 use mxlink::MessageResponseType;
 
-use crate::entity::{roomconfig::RoomSettings, MessageContext};
+use crate::entity::{MessageContext, roomconfig::RoomSettings};
-use crate::{strings, Bot};
+use crate::{Bot, strings};
 
 pub async fn handle_set<T>(
     bot: &Bot,

src/controller/cfg/room_config/handler.rs

@@ -1,9 +1,10 @@
 use mxlink::MessageResponseType;
 
 use crate::{
+    Bot,
     agent::{AgentPurpose, PublicIdentifier},
     entity::MessageContext,
-    strings, Bot,
+    strings,
 };
 
 use crate::entity::roomconfig::RoomConfigurationManager;

src/controller/cfg/status.rs

@@ -1,15 +1,16 @@
 use mxlink::MessageResponseType;
 
 use crate::{
+    Bot,
     agent::{
-        utils::get_effective_agent_for_purpose, AgentInstance, AgentPurpose, ControllerTrait,
+        AgentInstance, AgentPurpose, ControllerTrait, Manager as AgentManager, PublicIdentifier,
-        Manager as AgentManager, PublicIdentifier,
+        utils::get_effective_agent_for_purpose,
     },
     entity::{
-        roomconfig::{RoomConfig, RoomSettingsHandler},
         MessageContext, RoomConfigContext,
+        roomconfig::{RoomConfig, RoomSettingsHandler},
     },
-    strings, Bot,
+    strings,
 };
 
 pub async fn handle(bot: &Bot, message_context: &MessageContext) -> anyhow::Result<()> {
@@ -68,7 +69,7 @@ pub async fn handle(bot: &Bot, message_context: &MessageContext) -> anyhow::Resu
     );
     message.push_str("\n\n");
 
-    // Image Generation
+    // Image Creation
     message.push_str(
         &generate_image_generation_section(agent_manager, message_context.room_config_context())
             .await,
@@ -507,6 +508,35 @@ async fn generate_speech_to_text_section(
         flow_type_set_where,
     ));
 
+    // Msg Type For Non Threaded Only Transcribed Messages
+
+    let effective_msg_type_for_non_threaded_only_transcribed_messages =
+        room_config_context.speech_to_text_msg_type_for_non_threaded_only_transcribed_messages();
+    let room_config_msg_type_for_non_threaded_only_transcribed_messages = room_config_context
+        .room_config
+        .settings
+        .speech_to_text
+        .msg_type_for_non_threaded_only_transcribed_messages;
+    let global_config_msg_type_for_non_threaded_only_transcribed_messages = room_config_context
+        .global_config
+        .fallback_room_settings
+        .speech_to_text
+        .msg_type_for_non_threaded_only_transcribed_messages;
+
+    let msg_type_for_non_threaded_only_transcribed_messages_set_where =
+        if room_config_msg_type_for_non_threaded_only_transcribed_messages.is_some() {
+            strings::cfg::status_badge_set_in_room_config()
+        } else if global_config_msg_type_for_non_threaded_only_transcribed_messages.is_some() {
+            strings::cfg::status_badge_set_in_global_config()
+        } else {
+            strings::cfg::status_badge_using_hardcoded_default()
+        };
+
+    message.push_str(&strings::cfg::status_speech_to_text_entry_msg_type_for_non_threaded_only_transcribed_messages(
+        effective_msg_type_for_non_threaded_only_transcribed_messages,
+        msg_type_for_non_threaded_only_transcribed_messages_set_where,
+    ));
+
     // Language
 
     let effective_language = room_config_context.speech_to_text_language();

src/controller/chat_completion/mod.rs

@@ -1,30 +1,31 @@
-use mxlink::matrix_sdk::ruma::events::room::message::AudioMessageEventContent;
 use mxlink::matrix_sdk::ruma::OwnedEventId;
+use mxlink::matrix_sdk::ruma::events::room::message::AudioMessageEventContent;
 use mxlink::{MatrixLink, MessageResponseType};
 
 use tracing::Instrument;
 
-use crate::agent::provider::{
-    SpeechToTextParams, TextGenerationParams, TextGenerationPromptVariables,
-};
 use crate::agent::AgentInstance;
 use crate::agent::AgentPurpose;
 use crate::agent::ControllerTrait;
+use crate::agent::provider::{
+    SpeechToTextParams, TextGenerationParams, TextGenerationPromptVariables,
+};
 use crate::controller::utils::agent::get_effective_agent_for_purpose_or_complain;
 use crate::conversation::matrix::MatrixMessageProcessingParams;
-use crate::entity::roomconfig::{
-    SpeechToTextFlowType, TextToSpeechBotMessagesFlowType, TextToSpeechUserMessagesFlowType,
-};
 use crate::entity::MessagePayload;
+use crate::entity::roomconfig::{
+    SpeechToTextFlowType, SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages,
+    TextToSpeechBotMessagesFlowType, TextToSpeechUserMessagesFlowType,
+};
 use crate::strings;
 use crate::utils::text_to_speech::create_transcribed_message_text;
 use crate::{
+    Bot,
     conversation::{
         create_llm_conversation_for_matrix_reply_chain, create_llm_conversation_for_matrix_thread,
         matrix::create_list_of_bot_user_prefixes_to_strip,
     },
     entity::MessageContext,
-    Bot,
 };
 
 #[derive(Debug, PartialEq)]
@@ -38,6 +39,8 @@ pub enum ChatCompletionControllerType {
 
     Audio,
 
+    Image,
+
     ThreadMention,
     ReplyMention,
 }
@@ -71,21 +74,35 @@ pub async fn handle(
     if let MessagePayload::Audio(audio_content) = &message_context.payload() {
         original_message_is_audio = true;
 
-        let response_type = match speech_to_text_flow_type {
+        let (response_type, msg_type) = match speech_to_text_flow_type {
             SpeechToTextFlowType::Ignore => {
                 tracing::debug!("Intentionally ignoring audio message");
                 return Ok(());
             }
             SpeechToTextFlowType::TranscribeAndGenerateText => {
                 tracing::debug!("Will be transcribing and possibly generating text..");
-                MessageResponseType::InThread(message_context.thread_info().clone())
+                (
+                    MessageResponseType::InThread(message_context.thread_info().clone()),
+                    SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages::Notice,
+                )
             }
             SpeechToTextFlowType::OnlyTranscribe => {
                 tracing::debug!("Will only be transcribing audio to text..");
                 if message_context.thread_info().is_thread_root_only() {
-                    MessageResponseType::Reply(message_context.thread_info().root_event_id.clone())
+                    let msg_type = message_context
+                        .room_config_context()
+                        .speech_to_text_msg_type_for_non_threaded_only_transcribed_messages();
+                    (
+                        MessageResponseType::Reply(
+                            message_context.thread_info().root_event_id.clone(),
+                        ),
+                        msg_type,
+                    )
                 } else {
-                    MessageResponseType::InThread(message_context.thread_info().clone())
+                    (
+                        MessageResponseType::InThread(message_context.thread_info().clone()),
+                        SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages::Notice,
+                    )
                 }
             }
         };
@@ -94,8 +111,14 @@ pub async fn handle(
             _typing_notice_guard = Some(bot.start_typing_notice(message_context.room()).await);
         }
 
-        let Some(speech_to_text_created_event_id_result) =
+        let Some(speech_to_text_created_event_id_result) = handle_stage_speech_to_text(
-            handle_stage_speech_to_text(bot, message_context, audio_content, response_type).await
+            bot,
+            message_context,
+            audio_content,
+            response_type,
+            msg_type,
+        )
+        .await
         else {
             return Ok(());
         };
@@ -282,6 +305,7 @@ async fn handle_stage_speech_to_text(
     message_context: &MessageContext,
     audio_content: &AudioMessageEventContent,
     response_type: MessageResponseType,
+    msg_type: SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages,
 ) -> Option<OwnedEventId> {
     let agent = get_effective_agent_for_purpose_or_complain(
         bot,
@@ -302,7 +326,7 @@ async fn handle_stage_speech_to_text(
         .react_no_fail(
             message_context.room(),
             message_context.event_id().clone(),
-            AgentPurpose::SpeechToText.emoji().to_owned(),
+            strings::PROGRESS_INDICATOR_EMOJI.to_owned(),
         )
         .await;
 
@@ -312,6 +336,7 @@ async fn handle_stage_speech_to_text(
         &agent,
         audio_content,
         response_type.clone(),
+        msg_type,
     )
     .await;
 
@@ -393,7 +418,8 @@ async fn handle_stage_text_generation(
         ChatCompletionControllerType::TextCommand
         | ChatCompletionControllerType::TextMention
         | ChatCompletionControllerType::TextDirect
-        | ChatCompletionControllerType::Audio => {
+        | ChatCompletionControllerType::Audio
+        | ChatCompletionControllerType::Image => {
             Some(message_context.combined_admin_and_user_regexes())
         }
 
@@ -415,6 +441,7 @@ async fn handle_stage_text_generation(
         // When we're triggered via a reply mention, the context is the whole reply chain upward of the message that triggered us.
         ChatCompletionControllerType::ReplyMention => {
             create_llm_conversation_for_matrix_reply_chain(
+                &matrix_link,
                 &bot.room_event_fetcher().clone(),
                 message_context.room(),
                 message_context.thread_info().last_event_id.clone(),
@@ -426,7 +453,7 @@ async fn handle_stage_text_generation(
         // Everything else is happening in a thread, so the context is the whole thread.
         _ => {
             create_llm_conversation_for_matrix_thread(
-                matrix_link.clone(),
+                &matrix_link,
                 message_context.room(),
                 message_context.thread_info().root_event_id.clone(),
                 &params,
@@ -572,6 +599,7 @@ async fn handle_stage_speech_to_text_actual_transcribing(
     agent: &AgentInstance,
     audio_content: &AudioMessageEventContent,
     response_type: MessageResponseType,
+    msg_type: SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages,
 ) -> anyhow::Result<OwnedEventId> {
     let src = &audio_content.source;
 
@@ -626,9 +654,6 @@ async fn handle_stage_speech_to_text_actual_transcribing(
     //
     // When sending a bare reply, we'd better annotate the message with a 🦻 reaction instead,
     // to make it clear to users that it's a transcription.
-    //
-    // Regardless of how we post this message, it will be posted as a notice,
-    // which can indicate to the bot (for potential future text-generation purposes) that this message is not a bot message.
     let (transcribed_text, annotate_message_with_reaction) =
         if let MessageResponseType::InThread(_) = response_type {
             (
@@ -639,10 +664,22 @@ async fn handle_stage_speech_to_text_actual_transcribing(
             (speech_to_text_result.text, true)
         };
 
-    let result = bot
+    let result = match msg_type {
-        .messaging()
+        SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages::Text => {
-        .send_notice_markdown_no_fail(message_context.room(), transcribed_text, response_type)
+            bot.messaging()
-        .await;
+                .send_text_markdown_no_fail(message_context.room(), transcribed_text, response_type)
+                .await
+        }
+        SpeechToTextMessageTypeForNonThreadedOnlyTranscribedMessages::Notice => {
+            bot.messaging()
+                .send_notice_markdown_no_fail(
+                    message_context.room(),
+                    transcribed_text,
+                    response_type,
+                )
+                .await
+        }
+    };
 
     let event_id = result
         .map(|result| result.event_id)

src/controller/controller_type.rs

@@ -23,5 +23,6 @@ pub enum ControllerType {
     ChatCompletion(super::chat_completion::ChatCompletionControllerType),
 
     ImageGeneration(String),
+    ImageEdit(String),
     StickerGeneration(String),
 }

src/controller/determination/mod.rs

@@ -4,8 +4,8 @@ mod tests;
 use super::chat_completion::ChatCompletionControllerType;
 use crate::{
     entity::{
-        roomconfig::TextGenerationPrefixRequirementType, InteractionTrigger, MessageContext,
+        InteractionTrigger, MessageContext, MessagePayload,
-        MessagePayload,
+        roomconfig::TextGenerationPrefixRequirementType,
     },
     strings,
 };
@@ -36,6 +36,18 @@ pub fn determine_controller(
                 first_thread_message.is_mentioning_bot,
             )
         }
+        MessagePayload::Image(_image_message_content) => {
+            let prefix_requirement_type = message_context
+                .room_config_context()
+                .text_generation_prefix_requirement_type();
+
+            match prefix_requirement_type {
+                TextGenerationPrefixRequirementType::CommandPrefix => ControllerType::Ignore,
+                TextGenerationPrefixRequirementType::No => {
+                    ControllerType::ChatCompletion(ChatCompletionControllerType::Image)
+                }
+            }
+        }
         MessagePayload::Encrypted(thread_info) => {
             if thread_info.is_thread_root_only() {
                 ControllerType::Error(strings::error::message_is_encrypted().to_owned())
@@ -84,7 +96,7 @@ fn determine_text_controller(
     }
 
     if let Some(prompt) = text.strip_prefix(&format!("{command_prefix} image")) {
-        return ControllerType::ImageGeneration(prompt.trim().to_owned());
+        return super::image::determine_controller(prompt.trim());
     }
 
     if let Some(prompt) = text.strip_prefix(&format!("{command_prefix} sticker")) {

src/controller/determination/tests.rs

@@ -84,9 +84,17 @@ fn determine_text_controller() {
             expected: ControllerType::Config(controller::cfg::ConfigControllerType::Help),
         },
         TestCase {
-            name: "Image generation",
+            name: "Generic image command causes usage help",
             input: "!bai image Draw a cat!",
             is_mentioning_bot: false,
+            room_text_generation_prefix_requirement_type:
+                super::TextGenerationPrefixRequirementType::No,
+            expected: ControllerType::UsageHelp,
+        },
+        TestCase {
+            name: "Image generation",
+            input: "!bai image create Draw a cat!",
+            is_mentioning_bot: false,
             room_text_generation_prefix_requirement_type:
                 super::TextGenerationPrefixRequirementType::No,
             expected: ControllerType::ImageGeneration("Draw a cat!".to_owned()),
@@ -142,8 +150,7 @@ fn determine_text_controller() {
         // This test case is the same as the one above, just with a different prefix requirement setting.
         // We expect the same result.
         TestCase {
-            name:
+            name: "Regular message with bot mention triggers completion (command prefix requirement)",
-                "Regular message with bot mention triggers completion (command prefix requirement)",
             input: "Regular text goes here",
             is_mentioning_bot: true,
             room_text_generation_prefix_requirement_type:

src/controller/dispatching.rs

@@ -1,6 +1,6 @@
 use mxlink::MessageResponseType;
 
-use crate::{entity::MessageContext, strings, Bot};
+use crate::{Bot, entity::MessageContext, strings};
 
 use super::ControllerType;
 
@@ -77,6 +77,10 @@ pub async fn dispatch_controller(
             )
             .await
         }
+        ControllerType::ImageEdit(prompt) => {
+            super::image::edit::handle(bot, bot.matrix_link().clone(), message_context, prompt)
+                .await
+        }
         ControllerType::StickerGeneration(prompt) => {
             super::image::generation::handle_sticker(
                 bot,

src/controller/help/mod.rs

@@ -1,6 +1,6 @@
 use mxlink::MessageResponseType;
 
-use crate::{entity::MessageContext, strings, Bot};
+use crate::{Bot, entity::MessageContext, strings};
 
 pub async fn handle(bot: &Bot, message_context: &MessageContext) -> anyhow::Result<()> {
     let sender_can_manage_global_config = message_context.sender_can_manage_global_config();

src/controller/image/determination/mod.rs

@@ -0,0 +1,16 @@
+use crate::controller::ControllerType;
+mod tests;
+
+pub fn determine_controller(text: &str) -> ControllerType {
+    let text = text.trim();
+
+    if let Some(prompt) = text.strip_prefix("create") {
+        return ControllerType::ImageGeneration(prompt.trim().to_owned());
+    }
+
+    if let Some(prompt) = text.strip_prefix("edit") {
+        return ControllerType::ImageEdit(prompt.trim().to_owned());
+    }
+
+    ControllerType::UsageHelp
+}

src/controller/image/determination/tests.rs

@@ -0,0 +1,38 @@
+#[test]
+fn determine_controller() {
+    struct TestCase {
+        name: &'static str,
+        input: &'static str,
+        expected: super::ControllerType,
+    }
+
+    let test_cases = vec![
+        TestCase {
+            name: "Top-level is usage help",
+            input: "",
+            expected: super::ControllerType::UsageHelp,
+        },
+        TestCase {
+            name: "Top-level with some text is usage help",
+            input: "Some text",
+            expected: super::ControllerType::UsageHelp,
+        },
+        TestCase {
+            name: "Image generation triggered by create prefix",
+            input: "create Some prompt",
+            expected: super::ControllerType::ImageGeneration("Some prompt".to_owned()),
+        },
+        TestCase {
+            name: "Image edit triggered by edit prefix",
+            input: "edit Turn this into an anime-style image",
+            expected: super::ControllerType::ImageEdit(
+                "Turn this into an anime-style image".to_owned(),
+            ),
+        },
+    ];
+
+    for test_case in test_cases {
+        let result = super::determine_controller(test_case.input);
+        assert_eq!(result, test_case.expected, "Test case: {}", test_case.name);
+    }
+}

src/controller/image/edit.rs

@@ -0,0 +1,162 @@
+use mxlink::{MatrixLink, MessageResponseType};
+
+use tracing::Instrument;
+
+use crate::agent::AgentPurpose;
+use crate::agent::ControllerTrait;
+use crate::agent::provider::ImageEditParams;
+use crate::agent::provider::ImageSource;
+use crate::controller::utils::agent::get_effective_agent_for_purpose_or_complain;
+use crate::conversation::create_llm_conversation_for_matrix_thread;
+use crate::conversation::matrix::MatrixMessageProcessingParams;
+use crate::strings;
+use crate::utils::mime::get_file_extension;
+use crate::{Bot, entity::MessageContext};
+
+pub async fn handle(
+    bot: &Bot,
+    matrix_link: MatrixLink,
+    message_context: &MessageContext,
+    original_prompt: &str,
+) -> anyhow::Result<()> {
+    let response_type = MessageResponseType::InThread(message_context.thread_info().clone());
+
+    let Some(agent) = get_effective_agent_for_purpose_or_complain(
+        bot,
+        message_context,
+        AgentPurpose::ImageGeneration,
+        response_type.clone(),
+        true,
+    )
+    .await
+    else {
+        return Ok(());
+    };
+
+    if message_context.thread_info().is_thread_root_only() {
+        return send_guide(bot, message_context).await;
+    }
+
+    let _typing_notice_guard = bot.start_typing_notice(message_context.room()).await;
+
+    let params = MatrixMessageProcessingParams::new(
+        bot.user_id().to_owned(),
+        Some(message_context.combined_admin_and_user_regexes()),
+    );
+
+    let conversation = create_llm_conversation_for_matrix_thread(
+        &matrix_link,
+        message_context.room(),
+        message_context.thread_info().root_event_id.clone(),
+        &params,
+    )
+    .await?;
+
+    let prompt = if conversation.messages.len() >= 2 {
+        // Skip the first message, which contains the original prompt (which we already have)
+        let other_messages = conversation.messages.iter().skip(1).cloned().collect();
+
+        super::prompt::build(original_prompt, other_messages)
+    } else {
+        original_prompt.to_owned()
+    };
+
+    let got_go_signal = conversation.messages.iter().any(|message| {
+        if let crate::conversation::llm::MessageContent::Text(text) = &message.content {
+            text.to_lowercase() == "go"
+        } else {
+            false
+        }
+    });
+
+    let image_sources: Vec<ImageSource> = conversation
+        .messages
+        .iter()
+        .filter_map(|message| {
+            if let crate::conversation::llm::MessageContent::Image(image_content) = &message.content
+            {
+                Some(image_content.clone().into())
+            } else {
+                None
+            }
+        })
+        .collect();
+
+    if !got_go_signal || image_sources.is_empty() {
+        // We don't send the guide again here to avoid being annoying.
+        return Ok(());
+    }
+
+    let span = tracing::debug_span!("image_edit", agent_id = agent.identifier().as_string());
+
+    let result = agent
+        .controller()
+        .create_image_edit(&prompt, image_sources, ImageEditParams::default())
+        .instrument(span)
+        .await;
+
+    let response = match result {
+        Ok(response) => response,
+        Err(err) => {
+            tracing::warn!(
+                "Error in room {} while trying to generate image edit via agent {}: {:?}",
+                message_context.room_id(),
+                agent.identifier(),
+                err,
+            );
+
+            bot.messaging()
+                .send_error_markdown_no_fail(
+                    message_context.room(),
+                    &strings::agent::error_while_serving_purpose(
+                        agent.identifier(),
+                        &AgentPurpose::ImageGeneration,
+                        &err,
+                    ),
+                    response_type,
+                )
+                .await;
+
+            return Ok(());
+        }
+    };
+
+    let attachment_body_text = format!(
+        "generated-image-edit.{}",
+        get_file_extension(&response.mime_type)
+    );
+
+    let mut event_content = matrix_link
+        .media()
+        .upload_and_prepare_event_content(
+            message_context.room(),
+            &response.mime_type,
+            response.bytes,
+            &attachment_body_text,
+        )
+        .await
+        .map_err(|e| anyhow::anyhow!("Failed to upload and prepare event: {}", e))?;
+
+    matrix_link
+        .messaging()
+        .send_event(
+            message_context.room(),
+            &mut event_content,
+            response_type.clone(),
+        )
+        .await?;
+
+    Ok(())
+}
+
+async fn send_guide(bot: &Bot, message_context: &MessageContext) -> anyhow::Result<()> {
+    bot.messaging()
+        .send_text_markdown_no_fail(
+            message_context.room(),
+            strings::image_edit::guide_how_to_proceed(),
+            MessageResponseType::InThread(message_context.thread_info().clone()),
+        )
+        .await;
+
+    Ok(())
+}

src/controller/image/generation.rs

@@ -2,14 +2,15 @@ use mxlink::{MatrixLink, MessageResponseType};
 
 use tracing::Instrument;
 
-use crate::agent::provider::ImageGenerationParams;
 use crate::agent::AgentPurpose;
 use crate::agent::ControllerTrait;
+use crate::agent::provider::ImageGenerationParams;
 use crate::controller::utils::agent::get_effective_agent_for_purpose_or_complain;
 use crate::conversation::create_llm_conversation_for_matrix_thread;
 use crate::conversation::matrix::MatrixMessageProcessingParams;
 use crate::strings;
-use crate::{entity::MessageContext, Bot};
+use crate::utils::mime::get_file_extension;
+use crate::{Bot, entity::MessageContext};
 
 // We may make this configurable (per room, etc.) in the future, but for now it's hardcoded.
 const STICKER_SIZE: &str = "256x256";
@@ -42,7 +43,7 @@ pub async fn handle_image(
     );
 
     let conversation = create_llm_conversation_for_matrix_thread(
-        matrix_link.clone(),
+        &matrix_link,
         message_context.room(),
         message_context.thread_info().root_event_id.clone(),
         &params,
@@ -63,11 +64,37 @@ pub async fn handle_image(
         agent_id = agent.identifier().as_string()
     );
 
-    let response = agent
+    let result = agent
         .controller()
         .generate_image(&prompt, ImageGenerationParams::default())
         .instrument(span)
-        .await?;
+        .await;
+
+    let response = match result {
+        Ok(response) => response,
+        Err(err) => {
+            tracing::warn!(
+                "Error in room {} while trying to generate image via agent {}: {:?}",
+                message_context.room_id(),
+                agent.identifier(),
+                err,
+            );
+
+            bot.messaging()
+                .send_error_markdown_no_fail(
+                    message_context.room(),
+                    &strings::agent::error_while_serving_purpose(
+                        agent.identifier(),
+                        &AgentPurpose::ImageGeneration,
+                        &err,
+                    ),
+                    response_type,
+                )
+                .await;
+
+            return Ok(());
+        }
+    };
 
     let actual_prompt = response.revised_prompt.as_deref().unwrap_or(&prompt);
 
@@ -81,7 +108,10 @@ pub async fn handle_image(
             .await;
     }
 
-    let attachment_body_text = format!("Generated image based on: {}", actual_prompt);
+    let attachment_body_text = format!(
+        "generated-image.{}",
+        get_file_extension(&response.mime_type)
+    );
 
     let mut event_content = matrix_link
         .media()
@@ -151,13 +181,42 @@ pub async fn handle_sticker(
         .with_cheaper_model_switching_allowed(true)
         .with_cheaper_quality_switching_allowed(true);
 
-    let response = agent
+    let result = agent
         .controller()
         .generate_image(original_prompt, params)
         .instrument(span)
-        .await?;
+        .await;
 
-    let attachment_body_text = format!("Generated sticker image based on: {}", original_prompt);
+    let response = match result {
+        Ok(response) => response,
+        Err(err) => {
+            tracing::warn!(
+                "Error in room {} while trying to generate sticker via agent {}: {:?}",
+                message_context.room_id(),
+                agent.identifier(),
+                err,
+            );
+
+            bot.messaging()
+                .send_error_markdown_no_fail(
+                    message_context.room(),
+                    &strings::agent::error_while_serving_purpose(
+                        agent.identifier(),
+                        &AgentPurpose::ImageGeneration,
+                        &err,
+                    ),
+                    response_type,
+                )
+                .await;
+
+            return Ok(());
+        }
+    };
+
+    let attachment_body_text = format!(
+        "generated-sticker.{}",
+        get_file_extension(&response.mime_type)
+    );
 
     let mut event_content = matrix_link
         .media()

src/controller/image/mod.rs

@@ -1,2 +1,6 @@
+mod determination;
+pub mod edit;
 pub mod generation;
 mod prompt;
+
+pub use determination::determine_controller;

src/controller/image/prompt.rs

@@ -1,11 +1,11 @@
-use crate::conversation::llm::{Author, Message};
+use crate::conversation::llm::{Author, Message, MessageContent};
 
 /// Builds a prompt from the original prompt and other messages in the conversation.
 ///
 /// Only messages authored by the user are considered.
 ///
-/// Messages that say "Again" (regardless of casing) are ignored. They are considered special messages
+/// Messages that say "Again" or "Go" (regardless of casing) are ignored. They are considered special messages
-/// which trigger re-generation, but do not need to be included in the prompt criteria.
+/// which trigger re-generation and "start" respectively, and do not need to be included in the prompt criteria.
 pub fn build(original_prompt: &str, other_messages: Vec<Message>) -> String {
     let mut prompt = original_prompt.to_owned();
 
@@ -14,7 +14,11 @@ pub fn build(original_prompt: &str, other_messages: Vec<Message>) -> String {
         .into_iter()
         .filter(|message| {
             if let Author::User = message.author {
-                message.message_text.to_lowercase() != "again"
+                if let MessageContent::Text(text) = &message.content {
+                    text.to_lowercase() != "again" && text.to_lowercase() != "go"
+                } else {
+                    false
+                }
             } else {
                 false
             }
@@ -24,9 +28,9 @@ pub fn build(original_prompt: &str, other_messages: Vec<Message>) -> String {
     if !other_messages.is_empty() {
         prompt.push_str("\nOther criteria:");
         for message in other_messages {
-            prompt.push_str(
+            if let MessageContent::Text(text) = &message.content {
-                format!("\n- {}", message.message_text.replace("\n", ". ").as_str()).as_str(),
+                prompt.push_str(format!("\n- {}", text.replace("\n", ". ").as_str()).as_str());
-            );
+            }
         }
     }
 
@@ -36,7 +40,7 @@ pub fn build(original_prompt: &str, other_messages: Vec<Message>) -> String {
 #[cfg(test)]
 mod tests {
     use super::build;
-    use super::{Author, Message};
+    use super::{Author, Message, MessageContent};
 
     struct TestCase {
         original_prompt: &'static str,
@@ -60,7 +64,7 @@ mod tests {
                 original_prompt: "Generate a picture of a dog",
                 messages: vec![Message {
                     author: Author::User,
-                    message_text: "Must be blue".to_owned(),
+                    content: MessageContent::Text("Must be blue".to_owned()),
                     timestamp,
                 }],
                 expected_prompt: "Generate a picture of a dog\nOther criteria:\n- Must be blue",
@@ -68,46 +72,52 @@ mod tests {
             // Multiple complex user messages dispersed with assistant messages
             TestCase {
                 original_prompt: "Generate a picture of an elephant",
-                messages: vec![Message {
+                messages: vec![
-                    author: Author::User,
+                    Message {
-                    message_text: "Must be blue".to_owned(),
+                        author: Author::User,
-                    timestamp,
+                        content: MessageContent::Text("Must be blue".to_owned()),
-                },
+                        timestamp,
-                Message {
+                    },
-                    author: Author::Assistant,
+                    Message {
-                    message_text: "Whatever".to_owned(),
+                        author: Author::Assistant,
-                    timestamp,
+                        content: MessageContent::Text("Whatever".to_owned()),
-                },
+                        timestamp,
-                Message {
+                    },
-                    author: Author::User,
+                    Message {
-                    message_text: "Must be 3-legged.\nMust be flying.".to_owned(),
+                        author: Author::User,
-                    timestamp,
+                        content: MessageContent::Text(
-                }],
+                            "Must be 3-legged.\nMust be flying.".to_owned(),
+                        ),
+                        timestamp,
+                    },
+                ],
                 expected_prompt: "Generate a picture of an elephant\nOther criteria:\n- Must be blue\n- Must be 3-legged.. Must be flying.",
             },
             // "Again" is ignored.
             TestCase {
                 original_prompt: "Generate a picture of a grizzly bear",
-                messages: vec![Message {
+                messages: vec![
-                    author: Author::User,
+                    Message {
-                    message_text: "Must be blue".to_owned(),
+                        author: Author::User,
-                    timestamp,
+                        content: MessageContent::Text("Must be blue".to_owned()),
-                },
+                        timestamp,
-                Message {
+                    },
-                    author: Author::Assistant,
+                    Message {
-                    message_text: "Whatever".to_owned(),
+                        author: Author::Assistant,
-                    timestamp,
+                        content: MessageContent::Text("Whatever".to_owned()),
-                },
+                        timestamp,
-                Message {
+                    },
-                    author: Author::User,
+                    Message {
-                    message_text: "Again".to_owned(),
+                        author: Author::User,
-                    timestamp,
+                        content: MessageContent::Text("Again".to_owned()),
-                },
+                        timestamp,
-                Message {
+                    },
-                    author: Author::User,
+                    Message {
-                    message_text: "again".to_owned(),
+                        author: Author::User,
-                    timestamp,
+                        content: MessageContent::Text("again".to_owned()),
-                }],
+                        timestamp,
+                    },
+                ],
                 expected_prompt: "Generate a picture of a grizzly bear\nOther criteria:\n- Must be blue",
             },
         ];

src/controller/join/mod.rs

@@ -1,13 +1,21 @@
 use mxlink::MessageResponseType;
 
 use crate::entity::RoomConfigContext;
-use crate::{strings, Bot};
+use crate::{Bot, strings};
 
 pub async fn handle(
     bot: &Bot,
     room: &mxlink::matrix_sdk::Room,
     room_config_context: &RoomConfigContext,
 ) -> anyhow::Result<()> {
+    if !bot.post_join_self_introduction_enabled() {
+        tracing::debug!(
+            "Post-join self-introduction is disabled - not sending introduction message"
+        );
+
+        return Ok(());
+    }
+
     let agent_manager = bot.agent_manager();
 
     bot.messaging()

src/controller/provider/mod.rs

@@ -1,6 +1,6 @@
 use mxlink::MessageResponseType;
 
-use crate::{agent::AgentProvider, entity::MessageContext, strings, Bot};
+use crate::{Bot, agent::AgentProvider, entity::MessageContext, strings};
 
 use super::ControllerType;
 

src/controller/reaction/mod.rs

@@ -3,9 +3,9 @@ use std::ops::Deref;
 use mxlink::MatrixLink;
 
 use crate::{
+    Bot,
     agent::AgentPurpose,
     entity::{MessageContext, MessagePayload},
-    Bot,
 };
 
 mod text_to_speech;

src/controller/reaction/text_to_speech.rs

@@ -1,7 +1,7 @@
 use mxlink::{MatrixLink, MessageResponseType};
 
 use mxlink::matrix_sdk::ruma::{
-    events::room::message::TextMessageEventContent, OwnedEventId, OwnedUserId,
+    OwnedEventId, OwnedUserId, events::room::message::TextMessageEventContent,
 };
 
 use crate::entity::roomconfig::{
@@ -9,8 +9,8 @@ use crate::entity::roomconfig::{
 };
 
 use crate::{
-    agent::AgentPurpose, controller::utils::agent::get_effective_agent_for_purpose_or_complain,
+    Bot, agent::AgentPurpose,
-    entity::MessageContext, Bot,
+    controller::utils::agent::get_effective_agent_for_purpose_or_complain, entity::MessageContext,
 };
 
 pub(super) async fn handle(
@@ -34,7 +34,9 @@ pub(super) async fn handle(
         reacted_to_event_sender_id,
         matrix_link.user_id(),
     ) {
-        tracing::debug!("Ignoring request for on-demand text-to-speech (via reaction) due to room configuration");
+        tracing::debug!(
+            "Ignoring request for on-demand text-to-speech (via reaction) due to room configuration"
+        );
         return Ok(());
     }
 

src/controller/usage/mod.rs

@@ -1,6 +1,6 @@
 use mxlink::MessageResponseType;
 
-use crate::{entity::MessageContext, strings, Bot};
+use crate::{Bot, entity::MessageContext, strings};
 
 use super::ControllerType;
 

src/controller/utils/agent.rs

@@ -1,16 +1,17 @@
 use mxlink::MessageResponseType;
 
 use crate::{
+    Bot,
     agent::{
-        utils::{get_effective_agent_for_purpose, AgentForPurposeDeterminationError},
         AgentInstance, AgentPurpose,
+        utils::{AgentForPurposeDeterminationError, get_effective_agent_for_purpose},
     },
     entity::MessageContext,
-    strings, Bot,
+    strings,
 };
 
-pub async fn get_effective_agent_for_purpose_or_complain<'a>(
+pub async fn get_effective_agent_for_purpose_or_complain(
-    bot: &'a Bot,
+    bot: &Bot,
     message_context: &MessageContext,
     agent_purpose: AgentPurpose,
     response_type: MessageResponseType,

src/controller/utils/mod.rs

@@ -1,8 +1,8 @@
 use mxlink::MessageResponseType;
 
 use crate::{
-    entity::{MessageContext, MessagePayload},
     Bot,
+    entity::{MessageContext, MessagePayload},
 };
 
 pub mod agent;