Release 1.14.0

Document the new tools feature in README, features.md, and providers.md.
Update the `!bai providers` command output to show vision/tools support
consistently for all providers.

Based on #62 by @yeslayla which migrated the OpenAI provider to the
Responses API.

See: https://github.com/etkecc/baibot/pull/62

CHANGELOG.md

@@ -1,3 +1,55 @@
+# (2026-02-04) Version 1.14.0
+
+- (**Feature**) The `openai` provider now uses OpenAI's [Responses API](https://platform.openai.com/docs/api-reference/responses) (instead of the older Chat Completions API), adding support for [🛠️ built-in tools](./docs/features.md#️-built-in-tools-openai-only) (`web_search` and `code_interpreter`). These tools are **disabled by default** and can be enabled via the `text_generation.tools` configuration (see the [sample configuration](https://github.com/etkecc/baibot/blob/c70387b0c38d8d0f30bba2179a2a21a3710dbeaf/docs/sample-provider-configs/openai.yml#L12-L15)). To enable tools on an existing agent, you need to [update the agent](./docs/agents.md#updating-agents) to re-create it with the `text_generation.tools` section added and enable the tools you need. Thanks to [Layla Manley](https://github.com/yeslayla) for the contribution in [#62](https://github.com/etkecc/baibot/pull/62)!
+
+- (**Bugfix**) Fix sticker generation for newer GPT image models (`gpt-image-1`, `gpt-image-1-mini`, `gpt-image-1.5`) which don't support the previously hardcoded `256x256` size (minimum is `1024x1024`)
+
+- (**Internal Improvement**) Dependency updates
+
+
+# (2026-01-23) Version 1.13.0
+
+- (**Improvement**) Extend auto-switching to support cheaper models (`gpt-image-1-mini`) for `gpt-image-1` and `gpt-image-1.5` when generating stickers ([e0b4a40](https://github.com/etkecc/baibot/commit/e0b4a40))
+
+- (**Internal Improvement**) Upgrade Rust compiler (1.92.0 -> 1.93.0) ([691aeeb](https://github.com/etkecc/baibot/commit/691aeeb))
+
+- (**Internal Improvement**) Dependency updates
+
+
+# (2025-12-21) Version 1.12.0
+
+- (**Improvement**) Upgrade [async-openai](https://crates.io/crates/async-openai) (0.31.1 -> 0.32.2) and add support for OpenAI's `gpt-image-1.5` model ([08c689a](https://github.com/etkecc/baibot/commit/08c689a), [f7bf3d7](https://github.com/etkecc/baibot/commit/f7bf3d7))
+
+- (**Internal Improvement**) Dependency updates
+
+
+# (2025-12-15) Version 1.11.0
+
+- (**Feature**) Add support for custom avatars via file path and for keeping the already-set avatar (for those who wish to manage it by themselves via other means). See the [sample config](./etc/app/config.yml.dist) for details. ([062fbbb](https://github.com/etkecc/baibot/commit/062fbbb8ef9ad600db483a431c5c782402191023))
+
+- (**Internal Improvement**) Dependency updates ([99bde53](https://github.com/etkecc/baibot/commit/99bde53ef648a5a9086a96778fde4a9dbc1ede58))
+
+- (**Internal Improvement**) Documentation updates ([b3fd8e5](https://github.com/etkecc/baibot/commit/b3fd8e548f83fe46398ced4760d7e2bb7588c24d))
+
+- (**Internal Improvement**) Upgrade Rust compiler (1.91.1 -> 1.92.0) ([22906aa](https://github.com/etkecc/baibot/commit/22906aa2d3cae51815fad2560a545eaa69c247b6))
+
+
+# (2025-12-06) Version 1.10.0
+
+- (**Internal Improvement**) Dependency updates. This version is based on [mxlink](https://crates.io/crates/mxlink)@1.11.0 (which is based on the newly released [matrix-sdk](https://crates.io/crates/matrix-sdk)@[0.16.0](https://github.com/matrix-org/matrix-rust-sdk/releases/tag/matrix-sdk-0.16.0).
+
+# (2025-11-30) Version 1.9.0
+
+- (**Internal Improvement**) Upgrade [async-openai](https://crates.io/crates/async-openai) from our own etkecc fork (0.28.1-patched) to the official upstream version 0.31.1. This upgrade required some code adaptations to the new module structure, etc. While tested, regressions are possible.
+
+# (2025-11-28) Version 1.8.3
+
+- (**Improvement**) Add support for the `BAIBOT_PERSISTENCE_SESSION_ENCRYPTION_KEY` environment variable for configuring `persistence.session_encryption_key`
+
+- (**Improvement**) Add support for the `BAIBOT_USER_ENCRYPTION_RECOVERY_RESET_ALLOWED` environment variable for configuring `user.encryption.recovery_reset_allowed`
+
+- (**Internal Improvement**) Dependency updates.
+
 # (2025-11-20) Version 1.8.2
 
 - (**Internal Improvement**) Dependency and compiler updates (Rust 1.89.0 -> 1.91.1).

Cargo.toml

@@ -7,7 +7,7 @@ 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 = "1.8.2"
+version = "1.14.0"
 edition = "2024"
 
 [lib]
@@ -17,23 +17,24 @@ path = "src/lib.rs"
 [dependencies]
 anthropic = { git = "https://github.com/etkecc/anthropic-rs.git", branch = "fix-content-block-image" }
 anyhow = "1.0.*"
-async-openai = { git = "https://github.com/etkecc/async-openai", branch = "async-openai-v0.28.1-patched" }
+async-openai = { version = "0.32.3", features = ["audio", "chat-completion", "image", "responses"] }
 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.
 # We add the `native-tls` feature, because of https://github.com/etkecc/rust-mxlink/issues/1
-matrix-sdk = { version = "0.14.0", default-features = false, features = ["native-tls"] }
+matrix-sdk = { version = "0.16.0", default-features = false, features = ["native-tls"] }
+mime_guess = "2.0.*"
 mxidwc = "1.0.*"
-mxlink = ">=1.10.0"
+mxlink = ">=1.11.0"
 etke_openai_api_rust = "0.1.*"
 quick_cache = "0.6.*"
 regex = "1.12.*"
 serde = { version = "1.0.*", features = ["derive"], default-features = false }
 serde_json = "1.0.*"
 serde_yaml = "0.9.*"
-tempfile = "3.23.*"
+tempfile = "3.24.*"
 tiktoken-rs = { version = "0.9.*", default-features = false }
-tokio = { version = "1.48.*", features = ["rt", "rt-multi-thread", "macros"] }
+tokio = { version = "1.49.*", 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:1.91.1-slim-trixie AS build
+FROM docker.io/rust:1.93.0-slim-trixie 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:1.91.1-slim-trixie AS build
+FROM docker.io/rust:1.93.0-slim-trixie AS build
 
 RUN apt-get update && apt-get install -y build-essential pkg-config libssl-dev libsqlite3-dev
 

README.md

@@ -17,7 +17,7 @@ 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 (though certain models may "see" images as well)
+  - [💬 text-generation](./docs/features.md#-text-generation): communicating with you via text (though certain models may "see" images as well). The [OpenAI provider](./docs/providers.md#openai) also supports [🛠️ built-in tools](./docs/features.md#️-built-in-tools-openai-only) (web search, code interpreter)
   - [🦻 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#image-generation): creating and editing images based on instructions

docs/access.md

@@ -43,7 +43,8 @@ Administrators cannot be changed without adjusting the bot's configuration on th
 
 Room-local agent managers are users privileged to **create their own [agents](./agents.md)** (see `!bai agent`) in rooms.
 
-**⚠️ WARNING**: Letting regular users create agents which contact arbitrary network services **may be a security issue**.
+> [!WARNING]
+> Letting regular users create agents which contact arbitrary network services **may be a security issue**.
 
 The following commands are available:
 - **Show** the currently allowed users: `!bai access room-local-agent-managers`

docs/configuration/README.md

@@ -12,12 +12,15 @@ This file is created from the template found in [etc/app/config.yml.dist](../../
 
 Certain keys can be left unset, in which case [📝 hardcoded defaults](../../src/entity/cfg/defaults.rs) would be used.
 
-Each configuration key found in the YAML configuration can be overridden by setting an environment variable (dots should be replaced with `_`). Example:
+Some configuration keys found in the YAML configuration can be overridden by setting an environment variable (dots should be replaced with `_`). Example:
 
 - to override `command_prefix`, set an environment variable `BAIBOT_COMMAND_PREFIX`
 - to override `homeserver.server_name`, set an environment variable `BAIBOT_HOMESERVER_SERVER_NAME`
 
-The static configuration contains an `initial_global_config` key, which is used to populate the bot's global configuration (stored as [dynamic configuration](#dynamic-configuration)) the first time the bot starts. Modifying this subsequently will not have any effect. After initial global configuration creation, it's expected to be managed dynamically via chat commands.
+You can see the list of supported environment variables in the [🦀 src/entity/cfg/env.rs](../../src/entity/cfg/env.rs) file.
+
+> [!WARNING]
+> The static configuration contains an `initial_global_config` key, which is used to populate the bot's global configuration (stored as [dynamic configuration](#dynamic-configuration)) the first time the bot starts. Modifying this subsequently will not have any effect. After initial global configuration creation, it's expected to be managed dynamically via chat commands.
 
 
 ### Dynamic configuration

docs/features.md

@@ -40,6 +40,23 @@ You may also wish to see:
 - [📖 Usage / 💬 Text Generation](./usage.md#-text-generation) section for more details on how to use the bot for Text Generation in a room
 
 
+#### 🛠️ Built-in Tools (OpenAI only)
+
+
+
+The [OpenAI provider](./providers.md#openai) supports built-in tools that extend the model's capabilities:
+
+- [🔍 Web Search](https://platform.openai.com/docs/guides/tools-web-search) (`web_search`): allows the model to search the web for up-to-date information. [🖼️ Screenshot](./screenshots/text-generation-tools-web-search.webp)
+
+- [💻 Code Interpreter](https://platform.openai.com/docs/guides/tools-code-interpreter) (`code_interpreter`): allows the model to write and execute Python code in a sandbox
+
+These tools are **disabled by default** and need to be explicitly enabled in the agent's `text_generation.tools` configuration. See the [OpenAI sample configuration](https://github.com/etkecc/baibot/blob/c70387b0c38d8d0f30bba2179a2a21a3710dbeaf/docs/sample-provider-configs/openai.yml#L12-L15) for reference.
+
+To enable tools on an existing dynamically-created agent, you need to [update the agent](./agents.md#updating-agents) to re-create it with the `text_generation.tools` section added and enable the tools you need
+
+💡 **Note**: These tools run on OpenAI's infrastructure and may incur additional costs. Web search results include citations that are incorporated into the response.
+
+
 #### On-demand involvement
 
 In the following 2 cases, it's useful to involve the bot in conversations on-demand:

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) (no vision), [🖌️ 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) (incl. vision, incl. [🛠️ tools](./features.md#️-built-in-tools-openai-only)), [🖌️ 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) (incl. vision)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (incl. vision, no tools)
 - 🗲 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) (no vision), [🦻 speech-to-text](./features.md#-speech-to-text)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (no vision, no tools), [🦻 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) (no vision), [🗣️ 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, no tools), [🗣️ 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) (no vision)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (no vision, no tools)
 - 🗲 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) (no vision)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (no vision, no tools)
 - 🗲 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-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)
+- 🌟 Capabilities: [🖌️ image-generation](./features.md#️-image-creation), [💬 text-generation](./features.md#-text-generation) (incl. vision, incl. [🛠️ tools](./features.md#️-built-in-tools-openai-only)), [🗣️ 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`
@@ -137,7 +137,7 @@ 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-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)
+- 🌟 Capabilities: [🖌️ image-generation](./features.md#️-image-creation), [💬 text-generation](./features.md#-text-generation) (no vision, no tools), [🗣️ 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`
@@ -151,7 +151,7 @@ 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) (no vision)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (no vision, no tools)
 - 🗲 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`
@@ -165,7 +165,7 @@ 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) (no vision)
+- 🌟 Capabilities: [💬 text-generation](./features.md#-text-generation) (no vision, no tools)
 - 🗲 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/openai.yml

@@ -1,7 +1,7 @@
 base_url: https://api.openai.com/v1
 api_key: YOUR_API_KEY_HERE
 text_generation:
-  model_id: gpt-5.1
+  model_id: gpt-5.2
   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
   # Reasoning models need to use `max_completion_tokens` instead of `max_response_tokens`.
@@ -9,6 +9,10 @@ text_generation:
   max_response_tokens: null
   max_completion_tokens: 128000
   max_context_tokens: 400000
+  # Built-in tools
+  tools:
+    web_search: false
+    code_interpreter: false
 speech_to_text:
   model_id: whisper-1
 text_to_speech:
@@ -17,7 +21,7 @@ text_to_speech:
   speed: 1.0
   response_format: opus
 image_generation:
-  model_id: gpt-image-1
+  model_id: gpt-image-1.5
   style: null
   size: null
   quality: null

etc/app/config.yml.dist

@@ -11,6 +11,12 @@ user:
   # Leave empty to use the default (baibot).
   name: baibot
 
+  # An optional path to an image file to be used as a custom avatar image.
+  # - null or empty string: use the default avatar
+  # - "keep": don't touch the avatar, keep whatever is already set
+  # - any other value: path to a custom avatar image file
+  avatar: null
+
   encryption:
     # An optional passphrase to use for backing up and recovering the bot's encryption keys.
     # You can use any string here.
@@ -76,7 +82,7 @@ agents:
     #     base_url: https://api.openai.com/v1
     #     api_key: ""
     #     text_generation:
-    #       model_id: gpt-5.1
+    #       model_id: gpt-5.2
     #       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
     #       # Reasoning models need to use `max_completion_tokens` instead of `max_response_tokens`.
@@ -84,6 +90,10 @@ agents:
     #       max_response_tokens: null
     #       max_completion_tokens: 128000
     #       max_context_tokens: 400000
+    #       # Built-in tools
+    #       tools:
+    #         web_search: false
+    #         code_interpreter: false
     #     speech_to_text:
     #       model_id: whisper-1
     #     text_to_speech:
@@ -92,7 +102,7 @@ agents:
     #       speed: 1.0
     #       response_format: opus
     #     image_generation:
-    #       model_id: gpt-image-1
+    #       model_id: gpt-image-1.5
     #       style: null
     #       size: null
     #       quality: null

etc/services/core/compose.yml

@@ -14,7 +14,7 @@ services:
     - /etc/passwd:/etc/passwd:ro
 
   synapse:
-    image: ghcr.io/element-hq/synapse:v1.142.1
+    image: ghcr.io/element-hq/synapse:v1.146.0
     user: "${UID}:${GID}"
     restart: unless-stopped
     entrypoint: python
@@ -27,7 +27,7 @@ services:
     - ./synapse/media-store:/media-store
 
   element-web:
-    image: ghcr.io/element-hq/element-web:v1.12.4
+    image: ghcr.io/element-hq/element-web:v1.12.9
     user: "${UID}:${GID}"
     restart: unless-stopped
     environment:

etc/services/ollama/compose.yml

@@ -1,6 +1,6 @@
 services:
   ollama:
-    image: docker.io/ollama/ollama:0.13.0
+    image: docker.io/ollama/ollama:0.15.4
     restart: unless-stopped
     ports:
       - "${SERVICE_OLLAMA_BIND_PORT_HTTP}:11434"

src/agent/provider/entity/agent_provider.rs

@@ -69,6 +69,7 @@ impl AgentProvider {
                 models_list_url: Some("https://docs.anthropic.com/en/docs/about-claude/models"),
                 supported_purposes: vec![AgentPurpose::TextGeneration],
                 text_generation_supports_vision: true,
+                text_generation_supports_tools: false,
             },
             Self::Groq => AgentProviderInfo {
                 id: Self::Groq.to_static_str(),
@@ -80,11 +81,12 @@ impl AgentProvider {
                 models_list_url: Some("https://console.groq.com/docs/models"),
                 supported_purposes: vec![AgentPurpose::TextGeneration, AgentPurpose::SpeechToText],
                 text_generation_supports_vision: false,
+                text_generation_supports_tools: false,
             },
             Self::LocalAI => AgentProviderInfo {
                 id: Self::LocalAI.to_static_str(),
                 name: "LocalAI",
-                description: "LocalAI is the free, Open Source OpenAI alternative. LocalAI act as a drop-in replacement REST API that’s compatible with OpenAI API specifications for local inferencing. It allows you to run LLMs, generate images, audio (and not only) locally or on-prem with consumer grade hardware, supporting multiple model families and architectures.",
+                description: "LocalAI is the free, Open Source OpenAI alternative. LocalAI act as a drop-in replacement REST API that's compatible with OpenAI API specifications for local inferencing. It allows you to run LLMs, generate images, audio (and not only) locally or on-prem with consumer grade hardware, supporting multiple model families and architectures.",
                 homepage_url: Some("https://localai.io/"),
                 wiki_url: None,
                 sign_up_url: None,
@@ -95,6 +97,7 @@ impl AgentProvider {
                     AgentPurpose::SpeechToText,
                 ],
                 text_generation_supports_vision: false,
+                text_generation_supports_tools: false,
             },
             Self::Mistral => AgentProviderInfo {
                 id: Self::Mistral.to_static_str(),
@@ -106,6 +109,7 @@ impl AgentProvider {
                 models_list_url: Some("https://docs.mistral.ai/getting-started/models/"),
                 supported_purposes: vec![AgentPurpose::TextGeneration],
                 text_generation_supports_vision: false,
+                text_generation_supports_tools: false,
             },
             Self::Ollama => AgentProviderInfo {
                 id: Self::Ollama.to_static_str(),
@@ -117,6 +121,7 @@ impl AgentProvider {
                 models_list_url: Some("https://ollama.com/library"),
                 supported_purposes: vec![AgentPurpose::TextGeneration],
                 text_generation_supports_vision: false,
+                text_generation_supports_tools: false,
             },
             Self::OpenAI => AgentProviderInfo {
                 id: Self::OpenAI.to_static_str(),
@@ -133,6 +138,7 @@ impl AgentProvider {
                     AgentPurpose::SpeechToText,
                 ],
                 text_generation_supports_vision: true,
+                text_generation_supports_tools: true,
             },
             Self::OpenAICompat => AgentProviderInfo {
                 id: Self::OpenAICompat.to_static_str(),
@@ -149,6 +155,7 @@ impl AgentProvider {
                     AgentPurpose::SpeechToText,
                 ],
                 text_generation_supports_vision: false,
+                text_generation_supports_tools: false,
             },
             Self::OpenRouter => AgentProviderInfo {
                 id: Self::OpenRouter.to_static_str(),
@@ -160,6 +167,7 @@ impl AgentProvider {
                 models_list_url: Some("https://openrouter.ai/models"),
                 supported_purposes: vec![AgentPurpose::TextGeneration],
                 text_generation_supports_vision: false,
+                text_generation_supports_tools: false,
             },
             Self::TogetherAI => AgentProviderInfo {
                 id: Self::TogetherAI.to_static_str(),
@@ -171,6 +179,7 @@ impl AgentProvider {
                 models_list_url: Some("https://api.together.xyz/models"),
                 supported_purposes: vec![AgentPurpose::TextGeneration],
                 text_generation_supports_vision: false,
+                text_generation_supports_tools: false,
             },
         }
     }
@@ -192,4 +201,5 @@ pub struct AgentProviderInfo {
     pub models_list_url: Option<&'static str>,
     pub supported_purposes: Vec<AgentPurpose>,
     pub text_generation_supports_vision: bool,
+    pub text_generation_supports_tools: bool,
 }

src/agent/provider/entity/image.rs

@@ -2,7 +2,7 @@ use mxlink::mime;
 
 #[derive(Default)]
 pub struct ImageGenerationParams {
-    pub size_override: Option<String>,
+    pub smallest_size_possible: bool,
 
     pub cheaper_model_switching_allowed: bool,
 
@@ -10,8 +10,8 @@ pub struct ImageGenerationParams {
 }
 
 impl ImageGenerationParams {
-    pub fn with_size_override(mut self, value: Option<String>) -> Self {
+    pub fn with_smallest_size_possible(mut self, value: bool) -> Self {
-        self.size_override = value;
+        self.smallest_size_possible = value;
         self
     }
 
@@ -56,12 +56,8 @@ impl ImageSource {
     }
 }
 
-impl From<ImageSource> for async_openai::types::ImageInput {
+impl From<ImageSource> for async_openai::types::images::ImageInput {
     fn from(value: ImageSource) -> Self {
-        async_openai::types::ImageInput::from_vec_u8(
+        async_openai::types::images::ImageInput::from_vec_u8(value.filename, value.bytes)
-            value.filename,
-            value.bytes,
-            value.mime_type.to_string(),
-        )
     }
 }

src/agent/provider/openai/config.rs

@@ -1,6 +1,6 @@
 use serde::{Deserialize, Serialize};
 
-use super::OPENAI_IMAGE_MODEL_GPT_IMAGE_1;
+use super::OPENAI_IMAGE_MODEL_GPT_IMAGE_1_DOT_5;
 use crate::agent::{default_prompt, provider::ConfigTrait};
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -64,6 +64,9 @@ pub struct TextGenerationConfig {
 
     #[serde(default)]
     pub max_context_tokens: u32,
+
+    #[serde(default)]
+    pub tools: ToolsConfig,
 }
 
 impl Default for TextGenerationConfig {
@@ -75,12 +78,22 @@ impl Default for TextGenerationConfig {
             max_response_tokens: None,
             max_completion_tokens: Some(128_000),
             max_context_tokens: 400_000,
+            tools: ToolsConfig::default(),
         }
     }
 }
 
 fn default_text_model_id() -> String {
-    "gpt-5.1".to_owned()
+    "gpt-5.2".to_owned()
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize, Default)]
+pub struct ToolsConfig {
+    #[serde(default)]
+    pub web_search: bool,
+
+    #[serde(default)]
+    pub code_interpreter: bool,
 }
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -104,16 +117,16 @@ fn default_speech_to_text_model_id() -> String {
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct TextToSpeechConfig {
     #[serde(default = "default_text_to_speech_model_id")]
-    pub model_id: async_openai::types::SpeechModel,
+    pub model_id: async_openai::types::audio::SpeechModel,
 
     #[serde(default = "default_text_to_speech_voice")]
-    pub voice: async_openai::types::Voice,
+    pub voice: async_openai::types::audio::Voice,
 
     #[serde(default = "default_text_to_speech_speed")]
     pub speed: f32,
 
     #[serde(default = "default_text_to_speech_response_format")]
-    pub response_format: async_openai::types::SpeechResponseFormat,
+    pub response_format: async_openai::types::audio::SpeechResponseFormat,
 }
 
 impl Default for TextToSpeechConfig {
@@ -127,22 +140,22 @@ impl Default for TextToSpeechConfig {
     }
 }
 
-fn default_text_to_speech_model_id() -> async_openai::types::SpeechModel {
+fn default_text_to_speech_model_id() -> async_openai::types::audio::SpeechModel {
-    async_openai::types::SpeechModel::Tts1Hd
+    async_openai::types::audio::SpeechModel::Tts1Hd
 }
 
-fn default_text_to_speech_voice() -> async_openai::types::Voice {
+fn default_text_to_speech_voice() -> async_openai::types::audio::Voice {
-    async_openai::types::Voice::Onyx
+    async_openai::types::audio::Voice::Onyx
 }
 
 fn default_text_to_speech_speed() -> f32 {
     1.0
 }
 
-fn default_text_to_speech_response_format() -> async_openai::types::SpeechResponseFormat {
+fn default_text_to_speech_response_format() -> async_openai::types::audio::SpeechResponseFormat {
     // The API defaults to mp3, but we prefer Opus because it's smaller.
     // Our clients should all have support for it.
-    async_openai::types::SpeechResponseFormat::Opus
+    async_openai::types::audio::SpeechResponseFormat::Opus
 }
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -150,19 +163,19 @@ pub struct ImageGenerationConfig {
     pub model_id: String,
 
     #[serde(default = "default_image_style")]
-    pub style: Option<async_openai::types::ImageStyle>,
+    pub style: Option<async_openai::types::images::ImageStyle>,
 
     #[serde(default = "default_image_size")]
-    pub size: Option<async_openai::types::ImageSize>,
+    pub size: Option<async_openai::types::images::ImageSize>,
 
     #[serde(default = "default_image_quality")]
-    pub quality: Option<async_openai::types::ImageQuality>,
+    pub quality: Option<async_openai::types::images::ImageQuality>,
 }
 
 impl Default for ImageGenerationConfig {
     fn default() -> Self {
         Self {
-            model_id: OPENAI_IMAGE_MODEL_GPT_IMAGE_1.to_owned(),
+            model_id: OPENAI_IMAGE_MODEL_GPT_IMAGE_1_DOT_5.to_owned(),
             style: default_image_style(),
             size: default_image_size(),
             quality: default_image_quality(),
@@ -173,23 +186,28 @@ impl Default for ImageGenerationConfig {
 impl ImageGenerationConfig {
     pub fn model_id_as_openai_image_model(
         &self,
-    ) -> Result<async_openai::types::ImageModel, String> {
+    ) -> Result<async_openai::types::images::ImageModel, String> {
         match self.model_id.as_str() {
-            "dall-e-2" => Ok(async_openai::types::ImageModel::DallE2),
+            "dall-e-2" => Ok(async_openai::types::images::ImageModel::DallE2),
-            "dall-e-3" => Ok(async_openai::types::ImageModel::DallE3),
+            "dall-e-3" => Ok(async_openai::types::images::ImageModel::DallE3),
-            other => Ok(async_openai::types::ImageModel::Other(other.to_owned())),
+            "gpt-image-1" => Ok(async_openai::types::images::ImageModel::GptImage1),
+            "gpt-image-1.5" => Ok(async_openai::types::images::ImageModel::GptImage1dot5),
+            "gpt-image-1-mini" => Ok(async_openai::types::images::ImageModel::GptImage1Mini),
+            other => Ok(async_openai::types::images::ImageModel::Other(
+                other.to_owned(),
+            )),
         }
     }
 }
 
-fn default_image_style() -> Option<async_openai::types::ImageStyle> {
+fn default_image_style() -> Option<async_openai::types::images::ImageStyle> {
     None
 }
 
-fn default_image_size() -> Option<async_openai::types::ImageSize> {
+fn default_image_size() -> Option<async_openai::types::images::ImageSize> {
     None
 }
 
-fn default_image_quality() -> Option<async_openai::types::ImageQuality> {
+fn default_image_quality() -> Option<async_openai::types::images::ImageQuality> {
     None
 }

src/agent/provider/openai/controller.rs

@@ -4,9 +4,15 @@ use async_openai::{
     Client as OpenAIClient,
     config::OpenAIConfig,
     types::{
-        ChatCompletionRequestMessage, CreateChatCompletionRequestArgs, CreateImageEditRequestArgs,
+        audio::{AudioInput, CreateSpeechRequestArgs, CreateTranscriptionRequestArgs},
-        CreateImageRequestArgs, CreateSpeechRequestArgs, CreateTranscriptionRequestArgs,
+        responses::{
-        DallE2ImageSize, Image, ImageModel, ImageResponseFormat,
+            CodeInterpreterContainerAuto, CodeInterpreterTool, CodeInterpreterToolContainer,
+            CreateResponseArgs, OutputItem, OutputMessageContent, Tool, WebSearchTool,
+        },
+        images::{
+            CreateImageEditRequestArgs, CreateImageRequestArgs,
+            Image, ImageInput, ImageModel, ImageResponseFormat,
+        },
     },
 };
 
@@ -25,12 +31,9 @@ use crate::{
 use crate::{
     agent::{
         AgentPurpose,
-        provider::{
+        provider::entity::{
-            entity::{
+            ImageEditResult, ImageGenerationResult, ImageSource, PingResult,
-                ImageEditResult, ImageGenerationResult, ImageSource, PingResult,
+            TextToSpeechParams, TextToSpeechResult,
-                TextToSpeechParams, TextToSpeechResult,
-            },
-            openai::utils::convert_string_to_enum,
         },
     },
     strings,
@@ -38,8 +41,6 @@ use crate::{
 
 use super::config::Config;
 
-use super::OPENAI_IMAGE_MODEL_GPT_IMAGE_1;
-
 #[derive(Debug, Clone)]
 pub struct Controller {
     config: Config,
@@ -128,28 +129,44 @@ impl ControllerTrait for Controller {
             conversation_messages.insert(0, prompt_message);
         }
 
-        let openai_conversation_messages: Vec<ChatCompletionRequestMessage> =
+        let input = super::utils::convert_llm_messages_to_openai_response_input(conversation_messages);
-            super::utils::convert_llm_messages_to_openai_messages(conversation_messages);
 
-        let messages_count = openai_conversation_messages.len();
+        let messages_count = match &input {
+            async_openai::types::responses::InputParam::Items(items) => items.len(),
+            _ => 1,
+        };
 
         let temperature = params
             .temperature_override
             .unwrap_or(text_generation_config.temperature);
 
-        let mut request_builder = CreateChatCompletionRequestArgs::default();
+        let mut request_builder = CreateResponseArgs::default();
 
         request_builder
             .model(&text_generation_config.model_id)
             .temperature(temperature)
-            .messages(openai_conversation_messages);
+            .input(input);
 
-        if let Some(max_response_tokens) = text_generation_config.max_response_tokens {
+        let mut tools = Vec::new();
-            request_builder.max_tokens(max_response_tokens);
+        if text_generation_config.tools.web_search {
+            tools.push(Tool::WebSearch(WebSearchTool::default()));
+        }
+        if text_generation_config.tools.code_interpreter {
+            tools.push(Tool::CodeInterpreter(CodeInterpreterTool {
+                container: CodeInterpreterToolContainer::Auto(
+                    CodeInterpreterContainerAuto::default(),
+                ),
+            }));
         }
 
-        if let Some(max_completion_tokens) = text_generation_config.max_completion_tokens {
+        if !tools.is_empty() {
-            request_builder.max_completion_tokens(max_completion_tokens);
+            request_builder.tools(tools);
+        }
+
+        if let Some(max_response_tokens) = text_generation_config.max_response_tokens {
+            request_builder.max_output_tokens(max_response_tokens);
+        } else if let Some(max_completion_tokens) = text_generation_config.max_completion_tokens {
+            request_builder.max_output_tokens(max_completion_tokens);
         }
 
         let request = request_builder.build()?;
@@ -159,33 +176,31 @@ impl ControllerTrait for Controller {
                 model = format!("{:?}", request.model),
                 ?messages_count,
                 request = request_as_json,
-                "Sending OpenAI chat completion API request"
+                "Sending OpenAI response API request"
             );
         }
 
-        let response = self.client.chat().create(request).await?;
+        let response = self.client.responses().create(request).await?;
 
         tracing::trace!(
             ?response,
-            "Got response from the OpenAI chat completion API"
+            "Got response from the OpenAI response API"
         );
 
-        // We only request 1 result, so there should only be 1 choice.
+        for item in response.output {
-        if let Some(choice) = response.choices.into_iter().next() {
+            if let OutputItem::Message(message) = item {
-            match choice.message.content {
+                for content in message.content {
-                Some(text) => {
+                    if let OutputMessageContent::OutputText(text_content) = content {
-                    return Ok(TextGenerationResult { text });
+                        return Ok(TextGenerationResult {
-                }
+                            text: text_content.text,
-                None => {
+                        });
-                    return Err(anyhow::anyhow!(
+                    }
-                        "No content was found in the response choice from the OpenAI chat completion API"
-                    ));
                 }
             }
         }
 
         Err(anyhow::anyhow!(
-            "No response messages choices were returned from the OpenAI chat completion API"
+            "No response messages choices were returned from the OpenAI response API"
         ))
     }
 
@@ -209,11 +224,7 @@ impl ControllerTrait for Controller {
 
         let request = CreateTranscriptionRequestArgs::default()
             .model(&speech_to_text_config.model_id)
-            .file(async_openai::types::AudioInput::from_vec_u8(
+            .file(AudioInput::from_vec_u8(filename, media))
-                filename,
-                media,
-                mime_type.to_string(),
-            ))
             .language(language.clone())
             .build()?;
 
@@ -223,7 +234,7 @@ impl ControllerTrait for Controller {
             "Sending OpenAI speech-to-text API request"
         );
 
-        let response = self.client.audio().transcribe(request).await?;
+        let response = self.client.audio().transcription().create(request).await?;
 
         tracing::trace!(
             ?response,
@@ -255,10 +266,13 @@ impl ControllerTrait for Controller {
         let model = if params.cheaper_model_switching_allowed {
             // Switch to a cheaper model
             match original_model {
-                async_openai::types::ImageModel::DallE2 => async_openai::types::ImageModel::DallE2,
+                ImageModel::DallE2 => ImageModel::DallE2,
-                async_openai::types::ImageModel::DallE3 => async_openai::types::ImageModel::DallE2,
+                ImageModel::DallE3 => ImageModel::DallE2,
-                async_openai::types::ImageModel::Other(_) => {
+                ImageModel::GptImage1 => ImageModel::GptImage1Mini,
-                    async_openai::types::ImageModel::DallE2
+                ImageModel::GptImage1dot5 => ImageModel::GptImage1Mini,
+                ImageModel::GptImage1Mini => ImageModel::GptImage1Mini,
+                ImageModel::Other(_) => {
+                    ImageModel::DallE2
                 }
             }
         } else {
@@ -269,11 +283,24 @@ impl ControllerTrait for Controller {
             // Switch to a cheaper quality
             match &image_generation_config.quality {
                 Some(quality) => match quality {
-                    async_openai::types::ImageQuality::Standard => {
+                    async_openai::types::images::ImageQuality::Standard => {
-                        Some(async_openai::types::ImageQuality::Standard)
+                        Some(async_openai::types::images::ImageQuality::Standard)
                     }
-                    async_openai::types::ImageQuality::HD => {
+                    async_openai::types::images::ImageQuality::HD => {
-                        Some(async_openai::types::ImageQuality::Standard)
+                        Some(async_openai::types::images::ImageQuality::Standard)
+                    }
+                    // New quality levels - keep as-is or downgrade to Standard
+                    async_openai::types::images::ImageQuality::High => {
+                        Some(async_openai::types::images::ImageQuality::Standard)
+                    }
+                    async_openai::types::images::ImageQuality::Medium => {
+                        Some(async_openai::types::images::ImageQuality::Medium)
+                    }
+                    async_openai::types::images::ImageQuality::Low => {
+                        Some(async_openai::types::images::ImageQuality::Low)
+                    }
+                    async_openai::types::images::ImageQuality::Auto => {
+                        Some(async_openai::types::images::ImageQuality::Auto)
                     }
                 },
                 None => None,
@@ -282,20 +309,21 @@ impl ControllerTrait for Controller {
             image_generation_config.quality.clone()
         };
 
-        let size = params
+        let size = if params.smallest_size_possible {
-            .size_override
+            Some(get_sticker_size(&model))
-            .map(|s| convert_string_to_enum::<async_openai::types::ImageSize>(&s).unwrap())
+        } else {
-            .or(image_generation_config.size);
+            image_generation_config.size
+        };
 
         let response_format = match model.clone() {
             ImageModel::DallE2 => Some(ImageResponseFormat::B64Json),
             ImageModel::DallE3 => Some(ImageResponseFormat::B64Json),
-            ImageModel::Other(model_str) => match model_str.as_str() {
+            // gpt-image-1 only outputs base64 and we don't need to specify the response format.
-                // gpt-image-1 only outputs base64 and we don't need to specify the response format.
+            // In fact, specifying the response format results in an error.
-                // In fact, specifying the response format results in an error.
+            ImageModel::GptImage1 => None,
-                OPENAI_IMAGE_MODEL_GPT_IMAGE_1 => None,
+            ImageModel::GptImage1Mini => None,
-                _ => Some(ImageResponseFormat::B64Json),
+            ImageModel::GptImage1dot5 => None,
-            },
+            ImageModel::Other(_) => Some(ImageResponseFormat::B64Json),
         };
 
         let mut request_builder = CreateImageRequestArgs::default();
@@ -329,15 +357,15 @@ impl ControllerTrait for Controller {
             "Sending OpenAI image generation API request"
         );
 
-        let response = self.client.images().create(request).await?;
+        let response = self.client.images().generate(request).await?;
 
         if let Some(image) = response.data.into_iter().next() {
             match image.deref() {
-                async_openai::types::Image::B64Json {
+                Image::B64Json {
                     b64_json,
                     revised_prompt,
                 } => {
-                    let bytes = base64_decode(b64_json)?;
+                    let bytes = base64_decode(b64_json.as_ref())?;
 
                     return Ok(ImageGenerationResult {
                         bytes,
@@ -374,15 +402,15 @@ impl ControllerTrait for Controller {
             return Err(anyhow::anyhow!("No image sources provided"));
         }
 
-        let mut image_inputs = Vec::new();
+        let mut image_inputs: Vec<ImageInput> = 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::images::ImageSize::S256x256) => Some(async_openai::types::images::ImageSize::S256x256),
-            Some(async_openai::types::ImageSize::S512x512) => Some(DallE2ImageSize::S512x512),
+            Some(async_openai::types::images::ImageSize::S512x512) => Some(async_openai::types::images::ImageSize::S512x512),
-            Some(async_openai::types::ImageSize::S1024x1024) => Some(DallE2ImageSize::S1024x1024),
+            Some(async_openai::types::images::ImageSize::S1024x1024) => Some(async_openai::types::images::ImageSize::S1024x1024),
             _ => None,
         };
 
@@ -391,16 +419,18 @@ impl ControllerTrait for Controller {
             .map_err(|err| anyhow::anyhow!(err))?;
 
         let response_format = match model.clone() {
-            async_openai::types::ImageModel::DallE2 => {
+            ImageModel::DallE2 => {
-                Some(async_openai::types::ImageResponseFormat::B64Json)
+                Some(ImageResponseFormat::B64Json)
             }
-            async_openai::types::ImageModel::DallE3 => {
+            ImageModel::DallE3 => {
-                Some(async_openai::types::ImageResponseFormat::B64Json)
+                Some(ImageResponseFormat::B64Json)
             }
-            async_openai::types::ImageModel::Other(model_str) => match model_str.as_str() {
+            // gpt-image-1 only outputs base64 and we don't need to specify the response format.
-                OPENAI_IMAGE_MODEL_GPT_IMAGE_1 => None,
+            // In fact, specifying the response format results in an error.
-                _ => Some(async_openai::types::ImageResponseFormat::B64Json),
+            ImageModel::GptImage1 => None,
-            },
+            ImageModel::GptImage1Mini => None,
+            ImageModel::GptImage1dot5 => None,
+            ImageModel::Other(_) => Some(ImageResponseFormat::B64Json),
         };
 
         let mut request_builder = CreateImageEditRequestArgs::default();
@@ -429,12 +459,12 @@ impl ControllerTrait for Controller {
             "Sending OpenAI image edit API request"
         );
 
-        let response = self.client.images().create_edit(request).await?;
+        let response = self.client.images().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)?;
+                    let bytes = base64_decode(b64_json.as_ref())?;
                     return Ok(ImageEditResult {
                         bytes,
                         mime_type: mxlink::mime::IMAGE_PNG,
@@ -471,7 +501,7 @@ impl ControllerTrait for Controller {
 
         let voice = if let Some(voice_string) = params.voice_override {
             // This is a hacky way to construct a Voice enum from the string we have.
-            let voice: serde_json::Result<async_openai::types::Voice> =
+            let voice: serde_json::Result<async_openai::types::audio::Voice> =
                 serde_json::from_str(&format!("\"{}\"", voice_string));
             match voice {
                 Ok(voice) => voice,
@@ -511,7 +541,7 @@ impl ControllerTrait for Controller {
             "Sending OpenAI text-to-speech API request"
         );
 
-        let result = self.client.audio().speech(request).await?;
+        let result = self.client.audio().speech().create(request).await?;
 
         Ok(TextToSpeechResult {
             bytes: result.bytes.into(),
@@ -570,15 +600,15 @@ impl ControllerTrait for Controller {
 }
 
 fn response_format_to_mime_type(
-    response_format: &async_openai::types::SpeechResponseFormat,
+    response_format: &async_openai::types::audio::SpeechResponseFormat,
 ) -> Option<mxlink::mime::Mime> {
     let content_type = match response_format {
-        async_openai::types::SpeechResponseFormat::Mp3 => "audio/mp3".to_owned(),
+        async_openai::types::audio::SpeechResponseFormat::Mp3 => "audio/mp3".to_owned(),
-        async_openai::types::SpeechResponseFormat::Wav => "audio/wav".to_owned(),
+        async_openai::types::audio::SpeechResponseFormat::Wav => "audio/wav".to_owned(),
-        async_openai::types::SpeechResponseFormat::Opus => "audio/ogg".to_owned(),
+        async_openai::types::audio::SpeechResponseFormat::Opus => "audio/ogg".to_owned(),
-        async_openai::types::SpeechResponseFormat::Aac => "audio/aac".to_owned(),
+        async_openai::types::audio::SpeechResponseFormat::Aac => "audio/aac".to_owned(),
-        async_openai::types::SpeechResponseFormat::Flac => "audio/flac".to_owned(),
+        async_openai::types::audio::SpeechResponseFormat::Flac => "audio/flac".to_owned(),
-        async_openai::types::SpeechResponseFormat::Pcm => "audio/L8".to_owned(),
+        async_openai::types::audio::SpeechResponseFormat::Pcm => "audio/L8".to_owned(),
     };
 
     match content_type.parse() {
@@ -606,3 +636,17 @@ fn audio_mime_type_to_file_name(mime_type: &mxlink::mime::Mime) -> Option<String
 
     Some(format!("audio.{}", file_extension))
 }
+
+/// Returns the smallest supported size for stickers based on what the image model supports.
+fn get_sticker_size(model: &ImageModel) -> async_openai::types::images::ImageSize {
+    use async_openai::types::images::ImageSize;
+
+    match model {
+        ImageModel::DallE2 => ImageSize::S256x256,
+        ImageModel::DallE3 => ImageSize::S1024x1024,
+        ImageModel::GptImage1 => ImageSize::S1024x1024,
+        ImageModel::GptImage1Mini => ImageSize::S1024x1024,
+        ImageModel::GptImage1dot5 => ImageSize::S1024x1024,
+        ImageModel::Other(_) => ImageSize::S1024x1024,
+    }
+}

src/agent/provider/openai/mod.rs

@@ -16,7 +16,7 @@ use super::super::AgentInstantiationResult;
 use super::ConfigTrait;
 use super::controller::ControllerType;
 
-pub const OPENAI_IMAGE_MODEL_GPT_IMAGE_1: &str = "gpt-image-1";
+pub const OPENAI_IMAGE_MODEL_GPT_IMAGE_1_DOT_5: &str = "gpt-image-1.5";
 
 pub fn create_controller_from_yaml_value_config(
     agent_id: &str,

src/agent/provider/openai/utils.rs

@@ -1,8 +1,6 @@
-use async_openai::types::{
+use async_openai::types::responses::{
-    ChatCompletionRequestAssistantMessageArgs, ChatCompletionRequestMessage,
+    EasyInputContent, EasyInputMessage, ImageDetail, InputContent, InputImageContent, InputItem,
-    ChatCompletionRequestMessageContentPartImage, ChatCompletionRequestSystemMessageArgs,
+    InputParam, MessageType, Role,
-    ChatCompletionRequestUserMessageArgs, ChatCompletionRequestUserMessageContent,
-    ChatCompletionRequestUserMessageContentPart, ImageUrlArgs,
 };
 
 use crate::conversation::llm::{
@@ -10,93 +8,41 @@ use crate::conversation::llm::{
 };
 use crate::utils::base64::base64_encode;
 
-pub fn convert_llm_messages_to_openai_messages(
+pub fn convert_llm_messages_to_openai_response_input(
     conversation_messages: Vec<LLMMessage>,
-) -> Vec<ChatCompletionRequestMessage> {
+) -> InputParam {
-    let mut openai_conversation_messages: Vec<ChatCompletionRequestMessage> =
+    let mut items = Vec::with_capacity(conversation_messages.len());
-        Vec::with_capacity(conversation_messages.len());
 
     for message in conversation_messages {
-        let openai_message = convert_llm_message_to_openai_message(message);
+        let role = match message.author {
-        if let Some(openai_message) = openai_message {
+            LLMAuthor::Prompt => Role::System,
-            openai_conversation_messages.push(openai_message);
+            LLMAuthor::Assistant => Role::Assistant,
-        }
+            LLMAuthor::User => Role::User,
-    }
+        };
 
-    openai_conversation_messages
+        let content = match message.content {
-}
+            LLMMessageContent::Text(text) => EasyInputContent::Text(text),
+            LLMMessageContent::Image(image_details) => {
+                let image_url = format!(
+                    "data:{};base64,{}",
+                    image_details.mime,
+                    base64_encode(&image_details.data)
+                );
 
-fn convert_llm_message_to_openai_message(
+                EasyInputContent::ContentList(vec![InputContent::InputImage(InputImageContent {
-    llm_message: LLMMessage,
+                    image_url: Some(image_url),
-) -> Option<ChatCompletionRequestMessage> {
+                    detail: ImageDetail::Auto,
-    match &llm_message.content {
+                    file_id: None,
-        LLMMessageContent::Text(text) => Some(match llm_message.author {
+                })])
-            LLMAuthor::Prompt => ChatCompletionRequestSystemMessageArgs::default()
-                .content(text.clone())
-                .build()
-                .expect("Failed building OpenAI system message")
-                .into(),
-            LLMAuthor::Assistant => ChatCompletionRequestAssistantMessageArgs::default()
-                .content(text.clone())
-                .build()
-                .expect("Failed building OpenAI assistant message")
-                .into(),
-            LLMAuthor::User => ChatCompletionRequestUserMessageArgs::default()
-                .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
-                }
             }
-        }
+        };
-    }
-}
 
-pub(super) fn convert_string_to_enum<T>(value: &str) -> Result<T, String>
+        items.push(InputItem::EasyMessage(EasyInputMessage {
-where
+            r#type: MessageType::Message,
-    T: serde::de::DeserializeOwned,
+            role,
-{
+            content,
-    // This is a hacky way to construct an enum from the string we have.
+        }));
-    let enum_result: serde_json::Result<T> = serde_json::from_str(&format!("\"{}\"", value));
-    match enum_result {
-        Ok(enum_result) => Ok(enum_result),
-        Err(err) => {
-            tracing::debug!(?err, "Failed to parse into enum");
-
-            Err(format!("The value ({}) is not supported.", value))
-        }
     }
+
+    InputParam::Items(items)
 }

src/agent/provider/openai_compat/config.rs

@@ -95,6 +95,7 @@ impl TryInto<OpenAITextGenerationConfig> for TextGenerationConfig {
             max_response_tokens: self.max_response_tokens,
             max_completion_tokens: None,
             max_context_tokens: self.max_context_tokens,
+            tools: Default::default(),
         })
     }
 }
@@ -161,11 +162,11 @@ impl TryInto<OpenAITextToSpeechConfig> for TextToSpeechConfig {
     type Error = String;
 
     fn try_into(self) -> Result<OpenAITextToSpeechConfig, Self::Error> {
-        let model_id = convert_string_to_enum::<async_openai::types::SpeechModel>(&self.model_id)?;
+        let model_id = convert_string_to_enum::<async_openai::types::audio::SpeechModel>(&self.model_id)?;
 
-        let voice = convert_string_to_enum::<async_openai::types::Voice>(&self.voice)?;
+        let voice = convert_string_to_enum::<async_openai::types::audio::Voice>(&self.voice)?;
 
-        let response_format = convert_string_to_enum::<async_openai::types::SpeechResponseFormat>(
+        let response_format = convert_string_to_enum::<async_openai::types::audio::SpeechResponseFormat>(
             &self.response_format,
         )?;
 
@@ -224,7 +225,7 @@ impl TryInto<OpenAIImageGenerationConfig> for ImageGenerationConfig {
 
     fn try_into(self) -> Result<OpenAIImageGenerationConfig, Self::Error> {
         let size = if let Some(size) = &self.size {
-            Some(convert_string_to_enum::<async_openai::types::ImageSize>(
+            Some(convert_string_to_enum::<async_openai::types::images::ImageSize>(
                 size,
             )?)
         } else {
@@ -232,7 +233,7 @@ impl TryInto<OpenAIImageGenerationConfig> for ImageGenerationConfig {
         };
 
         let style = if let Some(style) = &self.style {
-            Some(convert_string_to_enum::<async_openai::types::ImageStyle>(
+            Some(convert_string_to_enum::<async_openai::types::images::ImageStyle>(
                 style,
             )?)
         } else {
@@ -240,7 +241,7 @@ impl TryInto<OpenAIImageGenerationConfig> for ImageGenerationConfig {
         };
 
         let quality = if let Some(quality) = &self.quality {
-            Some(convert_string_to_enum::<async_openai::types::ImageQuality>(
+            Some(convert_string_to_enum::<async_openai::types::images::ImageQuality>(
                 quality,
             )?)
         } else {

src/agent/provider/openai_compat/controller.rs

@@ -3,6 +3,8 @@ use etke_openai_api_rust::chat::{ChatApi, ChatBody};
 use etke_openai_api_rust::images::{ImagesApi, ImagesBody};
 use etke_openai_api_rust::{Auth, Message, OpenAI};
 
+const SMALLEST_IMAGE_SIZE: &str = "256x256";
+
 use super::super::ControllerTrait;
 use crate::utils::base64::base64_decode;
 use crate::{
@@ -303,9 +305,11 @@ impl ControllerTrait for Controller {
         // when they span multiple lines.
         let prompt = prompt.replace("\n", " ");
 
-        let size: Option<String> = params
+        let size: Option<String> = if params.smallest_size_possible {
-            .size_override
+            Some(SMALLEST_IMAGE_SIZE.to_owned())
-            .or_else(|| image_generation_config.size.clone());
+        } else {
+            image_generation_config.size.clone()
+        };
 
         let request = ImagesBody {
             model: Some(image_generation_config.model_id.to_owned()),

src/bot/implementation.rs

@@ -1,3 +1,4 @@
+use std::fs;
 use std::sync::Arc;
 use std::{future::Future, pin::Pin};
 
@@ -18,12 +19,13 @@ use mxlink::helpers::account_data_config::{
     RoomConfigManager as AccountDataRoomConfigManager,
 };
 use mxlink::helpers::encryption::Manager as EncryptionManager;
+use mxlink::mime::Mime;
 
 use crate::agent::Manager as AgentManager;
 use crate::entity::catch_up_marker::{
     CatchUpMarker, CatchUpMarkerManager, DelayedCatchUpMarkerManager,
 };
-use crate::entity::cfg::Config;
+use crate::entity::cfg::{Avatar, Config};
 use crate::entity::globalconfig::{GlobalConfig, GlobalConfigurationManager};
 use crate::entity::roomconfig::{RoomConfig, RoomConfigurationManager};
 
@@ -316,34 +318,72 @@ impl Bot {
             }
         }
 
-        let should_update_avatar = match &current_avatar_url {
+        let desired_avatar: Option<(Vec<u8>, Mime)> = match &self.inner.config.user.avatar {
-            Some(avatar_url) => {
+            Avatar::Keep => {
-                let request = MediaRequestParameters {
+                tracing::info!("Avatar configured to keep current, skipping avatar management");
-                    source: MediaSource::Plain(avatar_url.to_owned()),
+                None
-                    format: MediaFormat::File,
+            }
-                };
+            Avatar::Default => {
-
+                tracing::info!("Avatar configured to use default");
-                let content = media
+                Some((
-                    .get_media_content(&request, true)
+                    LOGO_BYTES.to_vec(),
-                    .await
+                    LOGO_MIME_TYPE
-                    .map_err(|e| anyhow::anyhow!("Failed fetching existing avatar: {:?}", e))?;
+                        .parse()
-
+                        .expect("Failed parsing mime type for logo"),
-                content.as_slice() != LOGO_BYTES
+                ))
+            }
+            Avatar::Custom(avatar_path) => {
+                tracing::info!(?avatar_path, "Avatar configured to use custom path");
+                let bytes = fs::read(avatar_path).map_err(|e| {
+                    anyhow::anyhow!("Failed reading avatar from {:?}: {:?}", avatar_path, e)
+                })?;
+                let mime = mime_guess::from_path(avatar_path).first_or_octet_stream();
+                tracing::debug!(?mime, bytes_len = bytes.len(), "Loaded custom avatar");
+                Some((bytes, mime))
             }
-            None => true,
         };
 
-        if should_update_avatar {
+        if let Some((desired_bytes, mime_type)) = desired_avatar {
-            tracing::info!("Updating avatar..");
+            let should_update_avatar = match &current_avatar_url {
+                Some(avatar_url) => {
+                    tracing::debug!(?avatar_url, "Fetching current avatar to compare");
+                    let request = MediaRequestParameters {
+                        source: MediaSource::Plain(avatar_url.to_owned()),
+                        format: MediaFormat::File,
+                    };
 
-            let mime_type = LOGO_MIME_TYPE
+                    let content = media
-                .parse()
+                        .get_media_content(&request, true)
-                .expect("Failed parsing mime type for logo");
+                        .await
+                        .map_err(|e| anyhow::anyhow!("Failed fetching existing avatar: {:?}", e))?;
 
-            account
+                    let needs_update = content.as_slice() != desired_bytes;
-                .upload_avatar(&mime_type, LOGO_BYTES.to_vec())
+
-                .await
+                    tracing::debug!(
-                .map_err(|e| anyhow::anyhow!("Failed uploading avatar: {:?}", e))?;
+                        current_bytes_len = content.len(),
+                        desired_bytes_len = desired_bytes.len(),
+                        ?needs_update,
+                        "Compared current and desired avatar"
+                    );
+
+                    needs_update
+                }
+                None => {
+                    tracing::debug!("No current avatar set, will upload");
+                    true
+                }
+            };
+
+            if should_update_avatar {
+                tracing::info!("Updating avatar..");
+                account
+                    .upload_avatar(&mime_type, desired_bytes)
+                    .await
+                    .map_err(|e| anyhow::anyhow!("Failed uploading avatar: {:?}", e))?;
+                tracing::info!("Avatar updated successfully");
+            } else {
+                tracing::debug!("Avatar already up to date, skipping upload");
+            }
         }
 
         Ok(())

src/bot/load_config.rs

@@ -5,7 +5,7 @@ use anyhow::anyhow;
 
 use crate::agent::AgentPurpose;
 
-pub use crate::entity::cfg::{Config, defaults as cfg_defaults, env as cfg_env};
+pub use crate::entity::cfg::{Avatar, 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)
@@ -33,7 +33,13 @@ pub fn load() -> anyhow::Result<Config> {
             cfg_env::BAIBOT_USER_ENCRYPTION_RECOVERY_PASSPHRASE => {
                 config.user.encryption.recovery_passphrase = Some(value);
             }
+            cfg_env::BAIBOT_USER_ENCRYPTION_RECOVERY_RESET_ALLOWED => {
+                config.user.encryption.recovery_reset_allowed = value.parse::<bool>()?;
+            }
             cfg_env::BAIBOT_USER_NAME => config.user.name = value,
+            cfg_env::BAIBOT_USER_AVATAR => {
+                config.user.avatar = Avatar::from_string(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>()?;
@@ -51,6 +57,9 @@ pub fn load() -> anyhow::Result<Config> {
             cfg_env::BAIBOT_PERSISTENCE_DATA_DIR_PATH => {
                 config.persistence.data_dir_path = Some(value);
             }
+            cfg_env::BAIBOT_PERSISTENCE_SESSION_ENCRYPTION_KEY => {
+                config.persistence.session_encryption_key = Some(value);
+            }
             cfg_env::BAIBOT_PERSISTENCE_CONFIG_ENCRYPTION_KEY => {
                 config.persistence.config_encryption_key = Some(value);
             }

src/controller/image/generation.rs

@@ -12,9 +12,6 @@ use crate::strings;
 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";
-
 pub async fn handle_image(
     bot: &Bot,
     matrix_link: MatrixLink,
@@ -177,7 +174,7 @@ pub async fn handle_sticker(
     );
 
     let params = ImageGenerationParams::default()
-        .with_size_override(Some(STICKER_SIZE.to_owned()))
+        .with_smallest_size_possible(true)
         .with_cheaper_model_switching_allowed(true)
         .with_cheaper_quality_switching_allowed(true);
 

src/entity/cfg/config.rs

@@ -1,7 +1,7 @@
 use std::path::PathBuf;
 
 use mxlink::helpers::encryption::EncryptionKey;
-use serde::{Deserialize, Serialize};
+use serde::{Deserialize, Deserializer, Serialize};
 
 use crate::{
     agent::{AgentDefinition, AgentPurpose, PublicIdentifier},
@@ -83,6 +83,52 @@ impl ConfigHomeserver {
     }
 }
 
+/// Configuration for the bot's avatar.
+///
+/// - `Default`: Use the built-in default avatar (null, empty string, or missing in config)
+/// - `Keep`: Don't touch the avatar, keep whatever is already set ("keep" in config)
+/// - `Custom(String)`: Use a custom avatar from the specified file path
+#[derive(Debug, Clone, PartialEq, Serialize)]
+pub enum Avatar {
+    /// Use the built-in default avatar
+    Default,
+    /// Keep the current avatar, don't change it
+    Keep,
+    /// Use a custom avatar from the specified file path
+    Custom(String),
+}
+
+impl Default for Avatar {
+    fn default() -> Self {
+        Avatar::Default
+    }
+}
+
+impl<'de> Deserialize<'de> for Avatar {
+    fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
+    where
+        D: Deserializer<'de>,
+    {
+        let value: Option<String> = Option::deserialize(deserializer)?;
+        Ok(match value {
+            None => Avatar::Default,
+            Some(s) => Avatar::from_string(s),
+        })
+    }
+}
+
+impl Avatar {
+    pub fn from_string(value: String) -> Self {
+        if value.is_empty() {
+            Avatar::Default
+        } else if value.eq_ignore_ascii_case("keep") {
+            Avatar::Keep
+        } else {
+            Avatar::Custom(value)
+        }
+    }
+}
+
 #[derive(Debug, Serialize, Deserialize)]
 pub struct ConfigUser {
     pub mxid_localpart: String,
@@ -93,6 +139,9 @@ pub struct ConfigUser {
 
     #[serde(default)]
     pub encryption: ConfigUserEncryption,
+
+    #[serde(default)]
+    pub avatar: Avatar,
 }
 
 impl ConfigUser {

src/entity/cfg/env.rs

@@ -6,8 +6,11 @@ pub const BAIBOT_HOMESERVER_URL: &str = "BAIBOT_HOMESERVER_URL";
 pub const BAIBOT_USER_MXID_LOCALPART: &str = "BAIBOT_USER_MXID_LOCALPART";
 pub const BAIBOT_USER_PASSWORD: &str = "BAIBOT_USER_PASSWORD";
 pub const BAIBOT_USER_NAME: &str = "BAIBOT_USER_NAME";
+pub const BAIBOT_USER_AVATAR: &str = "BAIBOT_USER_AVATAR";
 pub const BAIBOT_USER_ENCRYPTION_RECOVERY_PASSPHRASE: &str =
     "BAIBOT_USER_ENCRYPTION_RECOVERY_PASSPHRASE";
+pub const BAIBOT_USER_ENCRYPTION_RECOVERY_RESET_ALLOWED: &str =
+    "BAIBOT_USER_ENCRYPTION_RECOVERY_RESET_ALLOWED";
 
 pub const BAIBOT_COMMAND_PREFIX: &str = "BAIBOT_COMMAND_PREFIX";
 

src/entity/cfg/mod.rs

@@ -2,4 +2,4 @@ mod config;
 pub mod defaults;
 pub mod env;
 
-pub use config::Config;
+pub use config::{Avatar, Config};

src/strings/provider.rs

@@ -109,11 +109,21 @@ pub fn help_provider_details(id: &str, info: &AgentProviderInfo) -> String {
         let mut purpose_line = format!("{} {}", purpose.emoji(), purpose.as_str());
 
         if let AgentPurpose::TextGeneration = purpose {
+            let mut extras = vec![];
+
             if info.text_generation_supports_vision {
-                purpose_line = format!("{} ({})", purpose_line, "incl. vision");
+                extras.push("incl. vision");
             } else {
-                purpose_line = format!("{} ({})", purpose_line, "no vision");
+                extras.push("no vision");
             }
+
+            if info.text_generation_supports_tools {
+                extras.push("incl. tools");
+            } else {
+                extras.push("no tools");
+            }
+
+            purpose_line = format!("{} ({})", purpose_line, extras.join(", "));
         }
 
         capabilities.push(purpose_line);

src/strings/usage.rs

@@ -64,7 +64,7 @@ To create a sticker, send a command like `%command_prefix% sticker A huge bowl o
 
 The difference from **creating images** is that the bot will:
 
-- create a smaller-resolution image (`256x256`) - smaller/quicker, but still good enough for a sticker
+- create a smaller-resolution image (as small as the model allows) - smaller/quicker, but still good enough for a sticker
 - potentially switch to a different (cheaper or otherwise more suitable) model, if available
 - post the image directly to the room (as a reply to your message), without starting a threaded conversation
 "#;