Switch back to upstream async-openai

Once async-openai v0.31.0 gets released as a final version,
we'll be able top pull it from crates.io.

This does not quite compile yet, because of https://github.com/64bit/async-openai/issues/465

CHANGELOG.md

@@ -1,7 +1,3 @@
-# (2025-11-20) Version 1.8.2
-
-- (**Internal Improvement**) Dependency and compiler updates (Rust 1.89.0 -> 1.91.1).
-
 # (2025-09-12) Version 1.8.1
 
 - (**Internal Improvement**) Dependency updates.

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 = "1.8.1"
 edition = "2024"
 
 [lib]
@@ -17,7 +17,7 @@ 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 = { git = "https://github.com/64bit/async-openai", branch = "main" }
 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.
@@ -27,13 +27,13 @@ mxidwc = "1.0.*"
 mxlink = ">=1.10.0"
 etke_openai_api_rust = "0.1.*"
 quick_cache = "0.6.*"
-regex = "1.12.*"
+regex = "1.11.*"
 serde = { version = "1.0.*", features = ["derive"], default-features = false }
 serde_json = "1.0.*"
 serde_yaml = "0.9.*"
-tempfile = "3.23.*"
-tiktoken-rs = { version = "0.9.*", default-features = false }
-tokio = { version = "1.48.*", features = ["rt", "rt-multi-thread", "macros"] }
+tempfile = "3.21.*"
+tiktoken-rs = { version = "0.7.*", default-features = false }
+tokio = { version = "1.47.*", 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:1.90.0-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:1.90.0-slim-trixie AS build
 
 RUN apt-get update && apt-get install -y build-essential pkg-config libssl-dev libsqlite3-dev
 

docs/providers.md

@@ -125,7 +125,10 @@ For services which are not fully compatible with the OpenAI API, consider using
   - 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`
 
-💡 When creating an agent, the bot will show you an up-to-date sample configuration for this provider which looks [like this](./sample-provider-configs/openai.yml).
+💡 When creating an agent, the bot will show you an up-to-date sample configuration for this provider which:
+
+- in the general case looks [like this](./sample-provider-configs/openai.yml)
+- for the [o1](https://platform.openai.com/docs/models/o1) models needs to look [like this](./sample-provider-configs/openai-o1.yml)
 
 
 ### OpenAI Compatible

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

@@ -0,0 +1,24 @@
+base_url: https://api.openai.com/v1
+api_key: YOUR_API_KEY_HERE
+text_generation:
+  model_id: o1-mini
+  # o1 models do not support a system prompt
+  prompt: null
+  temperature: 1.0
+  # o1 models do not support max_response_tokens.
+  # They use `max_completion_tokens` as an alternative
+  max_response_tokens: null
+  max_completion_tokens: 16384
+  max_context_tokens: 128000
+speech_to_text:
+  model_id: whisper-1
+text_to_speech:
+  model_id: tts-1-hd
+  voice: onyx
+  speed: 1.0
+  response_format: opus
+image_generation:
+  model_id: gpt-image-1
+  style: null
+  size: null
+  quality: null

docs/sample-provider-configs/openai.yml

@@ -1,14 +1,11 @@
 base_url: https://api.openai.com/v1
 api_key: YOUR_API_KEY_HERE
 text_generation:
-  model_id: gpt-5.1
+  model_id: gpt-5
   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`.
-  # If you're dealing with a non-reasoning model, specify `max_response_tokens` and unset `max_completion_tokens`.
-  max_response_tokens: null
-  max_completion_tokens: 128000
-  max_context_tokens: 400000
+  max_response_tokens: 16384
+  max_context_tokens: 128000
 speech_to_text:
   model_id: whisper-1
 text_to_speech:

etc/app/config.yml.dist

@@ -76,12 +76,11 @@ agents:
     #     base_url: https://api.openai.com/v1
     #     api_key: ""
     #     text_generation:
-    #       model_id: gpt-5.1
+    #       model_id: gpt-5
     #       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: ~
     #       # Reasoning models need to use `max_completion_tokens` instead of `max_response_tokens`.
-    #       # If you're dealing with a non-reasoning model, specify `max_response_tokens` and unset `max_completion_tokens`.
-    #       max_response_tokens: null
     #       max_completion_tokens: 128000
     #       max_context_tokens: 400000
     #     speech_to_text:

etc/services/core/compose.yml

@@ -1,6 +1,6 @@
 services:
   postgres:
-    image: docker.io/postgres:18.1-alpine
+    image: docker.io/postgres:18.0-alpine
     user: ${UID}:${GID}
     restart: unless-stopped
     environment:
@@ -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:v1.140.0
     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:v1.12.2
     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:0.12.6
     restart: unless-stopped
     ports:
       - "${SERVICE_OLLAMA_BIND_PORT_HTTP}:11434"

src/agent/provider/entity/image.rs

@@ -58,10 +58,6 @@ impl ImageSource {
 
 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(),
-        )
+        async_openai::types::ImageInput::from_vec_u8(value.filename, value.bytes)
     }
 }

src/agent/provider/openai/config.rs

@@ -80,7 +80,7 @@ impl Default for TextGenerationConfig {
 }
 
 fn default_text_model_id() -> String {
-    "gpt-5.1".to_owned()
+    "gpt-5".to_owned()
 }
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -104,16 +104,16 @@ 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 +127,22 @@ impl Default for TextToSpeechConfig {
     }
 }
 
-fn default_text_to_speech_model_id() -> async_openai::types::SpeechModel {
-    async_openai::types::SpeechModel::Tts1Hd
+fn default_text_to_speech_model_id() -> async_openai::types::audio::SpeechModel {
+    async_openai::types::audio::SpeechModel::Tts1Hd
 }
 
-fn default_text_to_speech_voice() -> async_openai::types::Voice {
-    async_openai::types::Voice::Onyx
+fn default_text_to_speech_voice() -> async_openai::types::audio::Voice {
+    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)]

src/agent/provider/openai/controller.rs

@@ -5,8 +5,9 @@ use async_openai::{
     config::OpenAIConfig,
     types::{
         ChatCompletionRequestMessage, CreateChatCompletionRequestArgs, CreateImageEditRequestArgs,
-        CreateImageRequestArgs, CreateSpeechRequestArgs, CreateTranscriptionRequestArgs,
+        CreateImageRequestArgs,
         DallE2ImageSize, Image, ImageModel, ImageResponseFormat,
+        audio::{AudioInput, CreateSpeechRequestArgs, CreateTranscriptionRequestArgs},
     },
 };
 
@@ -209,11 +210,7 @@ impl ControllerTrait for Controller {
 
         let request = CreateTranscriptionRequestArgs::default()
             .model(&speech_to_text_config.model_id)
-            .file(async_openai::types::AudioInput::from_vec_u8(
-                filename,
-                media,
-                mime_type.to_string(),
-            ))
+            .file(AudioInput::from_vec_u8(filename, media))
             .language(language.clone())
             .build()?;
 
@@ -223,7 +220,7 @@ 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,
@@ -275,6 +272,19 @@ impl ControllerTrait for Controller {
                     async_openai::types::ImageQuality::HD => {
                         Some(async_openai::types::ImageQuality::Standard)
                     }
+                    // New quality levels - keep as-is or downgrade to Standard
+                    async_openai::types::ImageQuality::High => {
+                        Some(async_openai::types::ImageQuality::Standard)
+                    }
+                    async_openai::types::ImageQuality::Medium => {
+                        Some(async_openai::types::ImageQuality::Medium)
+                    }
+                    async_openai::types::ImageQuality::Low => {
+                        Some(async_openai::types::ImageQuality::Low)
+                    }
+                    async_openai::types::ImageQuality::Auto => {
+                        Some(async_openai::types::ImageQuality::Auto)
+                    }
                 },
                 None => None,
             }
@@ -471,7 +481,7 @@ 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 +521,7 @@ 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 +580,15 @@ 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::SpeechResponseFormat::Wav => "audio/wav".to_owned(),
-        async_openai::types::SpeechResponseFormat::Opus => "audio/ogg".to_owned(),
-        async_openai::types::SpeechResponseFormat::Aac => "audio/aac".to_owned(),
-        async_openai::types::SpeechResponseFormat::Flac => "audio/flac".to_owned(),
-        async_openai::types::SpeechResponseFormat::Pcm => "audio/L8".to_owned(),
+        async_openai::types::audio::SpeechResponseFormat::Mp3 => "audio/mp3".to_owned(),
+        async_openai::types::audio::SpeechResponseFormat::Wav => "audio/wav".to_owned(),
+        async_openai::types::audio::SpeechResponseFormat::Opus => "audio/ogg".to_owned(),
+        async_openai::types::audio::SpeechResponseFormat::Aac => "audio/aac".to_owned(),
+        async_openai::types::audio::SpeechResponseFormat::Flac => "audio/flac".to_owned(),
+        async_openai::types::audio::SpeechResponseFormat::Pcm => "audio/L8".to_owned(),
     };
 
     match content_type.parse() {

src/agent/provider/openai_compat/config.rs

@@ -161,11 +161,11 @@ 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,
         )?;