feat: LiquidAI audio plugin for LiveKit Agents #4656
Conversation
|
|
📝 WalkthroughWalkthroughAdds a new LiquidAI audio plugin package for LiveKit Agents: initializes and registers a LiquidAIPlugin at import, exposes STT/TTS and version, provides a module logger, and introduces packaging and README files. Changes
Sequence Diagram(s)sequenceDiagram
participant Client
participant STT as STT Engine
participant API as LiquidAI API
participant Audio as Audio Processing
Client->>STT: recognize(audio_buffer)
STT->>Audio: convert to WAV & base64
STT->>API: streaming /chat/completions (system + audio)
API-->>STT: stream text deltas
STT->>STT: assemble transcript
STT->>Client: SpeechEvent(FINAL_TRANSCRIPT)
sequenceDiagram
participant Client
participant TTS as TTS Engine
participant API as LiquidAI API
participant Audio as Audio Processing
participant Emitter as AudioEmitter
Client->>TTS: synthesize(text)
TTS->>API: streaming /chat/completions (system + text)
API-->>TTS: stream audio chunks (base64 float32)
TTS->>Audio: base64 decode & float32→int16
Audio->>Emitter: push PCM samples
TTS->>Emitter: flush
TTS->>Client: deliver ChunkedStream
Estimated Code Review Effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Suggested Reviewers
Poem
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing touches
🧪 Generate unit tests (beta)
📜 Recent review detailsConfiguration used: Organization UI Review profile: CHILL Plan: Pro 📒 Files selected for processing (1)
🧰 Additional context used📓 Path-based instructions (1)**/*.py📄 CodeRabbit inference engine (AGENTS.md)
Files:
🧠 Learnings (3)📓 Common learnings📚 Learning: 2026-01-16T07:44:56.353ZApplied to files:
📚 Learning: 2026-01-18T01:09:01.847ZApplied to files:
🧬 Code graph analysis (1)livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py (4)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (3)
🔇 Additional comments (5)
✏️ Tip: You can disable this entire section by setting Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Fix all issues with AI agents
In `@livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py`:
- Around line 120-128: The model parameter in the streaming call to
self._client.chat.completions.create is an empty string, causing HTTP 400;
change the call to pass the instance's model property (self.model) so the
correct model identifier (e.g., "LFM2.5-Audio") is sent; update the invocation
of self._client.chat.completions.create to replace model="" with
model=self.model and ensure any tests/configs set self.model appropriately.
In `@livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py`:
- Around line 143-151: The chat completion call is using an empty model string,
which will be rejected; update the call to pass the correct LiquidAI model
identifier by replacing model="" with the TTS instance's configured model (use
self.tts.model, which is "LFM2.5-Audio") in the
self._client.chat.completions.create invocation so the request uses the proper
model name.
In `@livekit-plugins/livekit-plugins-liquidai/pyproject.toml`:
- Line 25: Update the dependency list in pyproject.toml to require a newer
OpenAI client and to declare httpx explicitly: replace the existing dependencies
array entry that includes "openai>=1.0.0" with a bumped version (for example
"openai>=1.107.2" or a stable "openai>=2.0.0") and add "httpx" (e.g.,
"httpx>=0.24.0") alongside existing items; note that stt.py and tts.py import
httpx directly so adding it as an explicit dependency ensures installations
include it.
📜 Review details
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (6)
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/__init__.pylivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/log.pylivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.pylivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.pylivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/version.pylivekit-plugins/livekit-plugins-liquidai/pyproject.toml
🧰 Additional context used
📓 Path-based instructions (1)
**/*.py
📄 CodeRabbit inference engine (AGENTS.md)
**/*.py: Format code with ruff
Run ruff linter and auto-fix issues
Run mypy type checker in strict mode
Maintain line length of 100 characters maximum
Ensure Python 3.9+ compatibility
Use Google-style docstrings
Files:
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/version.pylivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/log.pylivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/__init__.pylivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.pylivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py
🧠 Learnings (3)
📓 Common learnings
Learnt from: CR
Repo: livekit/agents PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-01-16T07:44:56.353Z
Learning: Implement Model Interface Pattern for STT, TTS, LLM, and Realtime models with provider-agnostic interfaces, fallback adapters for resilience, and stream adapters for different streaming patterns
Learnt from: davidzhao
Repo: livekit/agents PR: 4548
File: livekit-plugins/livekit-plugins-openai/livekit/plugins/openai/responses/llm.py:77-81
Timestamp: 2026-01-18T01:09:01.847Z
Learning: In the OpenAI responses LLM (`livekit-plugins-openai/livekit/plugins/openai/responses/llm.py`), reasoning effort defaults are intentionally set lower than OpenAI's API defaults for voice interactions: "none" for gpt-5.1/gpt-5.2 and "minimal" for other reasoning-capable models like gpt-5, to avoid enabling reasoning by default in voice contexts.
Learnt from: CR
Repo: livekit/agents PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-01-16T07:44:56.353Z
Learning: Follow the Plugin System pattern where plugins in livekit-plugins/ are separate packages registered via the Plugin base class
📚 Learning: 2026-01-16T07:44:56.353Z
Learnt from: CR
Repo: livekit/agents PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-01-16T07:44:56.353Z
Learning: Follow the Plugin System pattern where plugins in livekit-plugins/ are separate packages registered via the Plugin base class
Applied to files:
livekit-plugins/livekit-plugins-liquidai/pyproject.toml
📚 Learning: 2026-01-16T07:44:56.353Z
Learnt from: CR
Repo: livekit/agents PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-01-16T07:44:56.353Z
Learning: Implement Model Interface Pattern for STT, TTS, LLM, and Realtime models with provider-agnostic interfaces, fallback adapters for resilience, and stream adapters for different streaming patterns
Applied to files:
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py
🧬 Code graph analysis (3)
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/__init__.py (3)
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py (1)
STT(28-156)livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py (1)
TTS(36-114)livekit-agents/livekit/agents/plugin.py (2)
Plugin(13-56)register_plugin(31-36)
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py (4)
livekit-agents/livekit/agents/_exceptions.py (2)
APIConnectionError(84-88)APITimeoutError(91-95)livekit-agents/livekit/agents/types.py (1)
APIConnectOptions(54-88)livekit-agents/livekit/agents/utils/misc.py (1)
is_given(25-26)livekit-agents/livekit/agents/tts/tts.py (1)
TTSCapabilities(47-51)
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py (3)
livekit-agents/livekit/agents/_exceptions.py (2)
APIConnectionError(84-88)APITimeoutError(91-95)livekit-agents/livekit/agents/stt/stt.py (4)
SpeechEventType(32-49)STTCapabilities(78-84)SpeechEvent(70-74)SpeechData(53-61)livekit-agents/livekit/agents/utils/misc.py (1)
is_given(25-26)
🔇 Additional comments (15)
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/version.py (1)
1-1: LGTM. Version constant is straightforward and in the right place.livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/log.py (1)
1-3: LGTM. Logger setup is minimal and clear.livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/__init__.py (2)
31-33: LGTM. Plugin initialization wires required metadata cleanly.
36-45: LGTM. Registration and doc-visibility pruning are straightforward.livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py (5)
17-25: LGTM. Defaults and options container are clear.
31-70: LGTM. Constructor sets capabilities and defaults cleanly.
80-89: LGTM. Option updates are guarded appropriately.
149-153: LGTM. SpeechEvent mapping is concise and correct.
155-156: LGTM. Client shutdown is handled.livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py (6)
21-34: LGTM. Defaults and options container are clean.
39-78: LGTM. Constructor wiring is clear.
88-97: LGTM. Option updates are straightforward.
99-111: LGTM. ChunkedStream construction is clean.
113-114: LGTM. Client shutdown is handled.
120-132: LGTM. Stream wrapper initialization is tidy.
✏️ Tip: You can disable this entire section by setting review_details to false in your review settings.
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py
Show resolved
Hide resolved
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py
Show resolved
Hide resolved
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Fix all issues with AI agents
In `@livekit-plugins/livekit-plugins-liquidai/README.md`:
- Line 5: Convert the bare URL in the README.md to a Markdown link by replacing
the plain URL "https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B-GGUF" with a
link label and URL format, e.g. use a descriptive label like
[LFM2.5-Audio-1.5B-GGUF](https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B-GGUF)
so the linter warning is resolved and the link is more readable.
- Line 3: Update the README description string "LFM2.5-Audio family of SST from
LiquidAI." to use the correct abbreviation "STT" (i.e., change "SST" -> "STT");
search the README.md for any other occurrences of "SST" and replace them with
"STT" so the phrase "LFM2.5-Audio family of STT from LiquidAI" is consistent.
- Around line 15-20: The README's "Start audio server" section uses the external
binary llama-liquid-audio-server (and CKPT paths) without guidance; update the
docs to state that llama-liquid-audio-server is provided by LiquidAI, link to
the official repository or release page, add brief build/install steps (git
clone, build prerequisites like Rust/CMake/GCC or where to download prebuilt
binaries), and list system prerequisites (OS, GPU/CPU, CUDA/CUDNN or AVX
requirements, disk for model checkpoints) plus an example of placing model files
under the CKPT directory so the existing command will work.
📜 Review details
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
livekit-plugins/livekit-plugins-liquidai/README.md
🧰 Additional context used
🧠 Learnings (2)
📓 Common learnings
Learnt from: CR
Repo: livekit/agents PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-01-16T07:44:56.353Z
Learning: Follow the Plugin System pattern where plugins in livekit-plugins/ are separate packages registered via the Plugin base class
📚 Learning: 2026-01-16T07:44:56.353Z
Learnt from: CR
Repo: livekit/agents PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-01-16T07:44:56.353Z
Learning: Follow the Plugin System pattern where plugins in livekit-plugins/ are separate packages registered via the Plugin base class
Applied to files:
livekit-plugins/livekit-plugins-liquidai/README.md
🪛 markdownlint-cli2 (0.20.0)
livekit-plugins/livekit-plugins-liquidai/README.md
[warning] 5-5: Bare URL used
(MD034, no-bare-urls)
🔇 Additional comments (1)
livekit-plugins/livekit-plugins-liquidai/README.md (1)
7-11: LGTM!The installation instructions are clear and follow standard conventions.
✏️ Tip: You can disable this entire section by setting review_details to false in your review settings.
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Fix all issues with AI agents
In `@livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py`:
- Around line 88-97: The voice option is stored but never applied; when
update_options receives voice (self._opts.voice) you must fold it into the
effective system prompt used for synthesis (self._opts.system_prompt) so the API
receives the voice directive. Modify update_options (and the request assembly
code that currently reads self._opts.system_prompt) to compute a combined prompt
(e.g., prepend or append a short "Voice: {voice}" instruction when
self._opts.voice is given) and use that combined prompt in the API request
payload instead of using system_prompt alone; ensure the same merging logic is
used wherever the request body is constructed so update_options(voice=...)
actually changes synthesis.
🧹 Nitpick comments (3)
livekit-plugins/livekit-plugins-liquidai/README.md (3)
3-3: Be more specific about the model family.The phrase "Audio family" is vague. Since the plugin specifically supports the LFM2.5-Audio models (as indicated in the title and HuggingFace link), clarify this in the description.
📝 Suggested improvement
-Support for the Audio family of SST/TTS from LiquidAI. +Support for the LFM2.5-Audio family of SST/TTS from LiquidAI.
18-21: Improve command readability with line continuation.The server startup command is very long and difficult to read on a single line. Breaking it into multiple lines using bash line continuation would make it easier for users to understand the command structure and parameters.
♻️ Suggested improvement
-```bash -export CKPT=/path/to/LFM2.5-Audio-1.5B-GGUF -./llama-liquid-audio-server -m $CKPT/LFM2.5-Audio-1.5B-Q4_0.gguf -mm $CKPT/mmproj-LFM2.5-Audio-1.5B-Q4_0.gguf -mv $CKPT/vocoder-LFM2.5-Audio-1.5B-Q4_0.gguf --tts-speaker-file $CKPT/tokenizer-LFM2.5-Audio-1.5B-Q4_0.gguf -``` +```bash +export CKPT=/path/to/LFM2.5-Audio-1.5B-GGUF +./llama-liquid-audio-server \ + -m $CKPT/LFM2.5-Audio-1.5B-Q4_0.gguf \ + -mm $CKPT/mmproj-LFM2.5-Audio-1.5B-Q4_0.gguf \ + -mv $CKPT/vocoder-LFM2.5-Audio-1.5B-Q4_0.gguf \ + --tts-speaker-file $CKPT/tokenizer-LFM2.5-Audio-1.5B-Q4_0.gguf +``` </details> --- `1-21`: **Consider adding usage examples.** The README would be more helpful with a brief usage example showing how to instantiate and use the STT and TTS classes in LiveKit Agents code. This would help users get started quickly after installation. <details> <summary>💡 Example structure</summary> Consider adding a section like: ```markdown ## Usage ```python from livekit import agents from livekit.plugins import liquidai # Speech-to-Text stt = liquidai.STT() # Text-to-Speech tts = liquidai.TTS()</details> </blockquote></details> </blockquote></details> <details> <summary>📜 Review details</summary> **Configuration used**: Organization UI **Review profile**: CHILL **Plan**: Pro <details> <summary>📥 Commits</summary> Reviewing files that changed from the base of the PR and between 7ac3560743b7c2b89c0e713139cbe16eb14bb932 and 810d1010e5cfc39a0e5b50277836ef541274b1e1. </details> <details> <summary>📒 Files selected for processing (2)</summary> * `livekit-plugins/livekit-plugins-liquidai/README.md` * `livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py` </details> <details> <summary>🧰 Additional context used</summary> <details> <summary>📓 Path-based instructions (1)</summary> <details> <summary>**/*.py</summary> **📄 CodeRabbit inference engine (AGENTS.md)** > `**/*.py`: Format code with ruff > Run ruff linter and auto-fix issues > Run mypy type checker in strict mode > Maintain line length of 100 characters maximum > Ensure Python 3.9+ compatibility > Use Google-style docstrings Files: - `livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py` </details> </details><details> <summary>🧠 Learnings (2)</summary> <details> <summary>📓 Common learnings</summary>Learnt from: CR
Repo: livekit/agents PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-01-16T07:44:56.353Z
Learning: Follow the Plugin System pattern where plugins in livekit-plugins/ are separate packages registered via the Plugin base class</details> <details> <summary>📚 Learning: 2026-01-18T01:09:01.847Z</summary>Learnt from: davidzhao
Repo: livekit/agents PR: 4548
File: livekit-plugins/livekit-plugins-openai/livekit/plugins/openai/responses/llm.py:77-81
Timestamp: 2026-01-18T01:09:01.847Z
Learning: In the OpenAI responses LLM (livekit-plugins-openai/livekit/plugins/openai/responses/llm.py), reasoning effort defaults are intentionally set lower than OpenAI's API defaults for voice interactions: "none" for gpt-5.1/gpt-5.2 and "minimal" for other reasoning-capable models like gpt-5, to avoid enabling reasoning by default in voice contexts.**Applied to files:** - `livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py` </details> </details> </details> <details> <summary>⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)</summary> * GitHub Check: unit-tests </details> <details> <summary>🔇 Additional comments (5)</summary><blockquote> <details> <summary>livekit-plugins/livekit-plugins-liquidai/README.md (1)</summary><blockquote> `8-12`: **LGTM!** The installation instructions are clear and follow standard Python package conventions. </blockquote></details> <details> <summary>livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py (4)</summary><blockquote> `21-34`: **Defaults/options setup looks good.** Clear defaults and a simple options container keep the surface area clean. --- `99-114`: **synthesize/aclose wiring looks solid.** Clean construction of the stream object and proper client shutdown. --- `120-132`: **ChunkedStream initialization is clean.** State setup and base initialization are straightforward. --- `56-60`: No changes needed. The `streaming=False` flag is correct. The `TTSCapabilities.streaming` flag indicates whether the TTS accepts streaming text input (incremental text chunks), not whether it outputs audio in chunks. LiquidAI's `synthesize()` requires the full text upfront and does not implement a `stream()` method for accepting text chunks, making `streaming=False` appropriate. This follows the standard pattern used by OpenAI, Azure, Google, and other non-streaming TTS implementations. The framework automatically wraps non-streaming TTS with `StreamAdapter` in the agent pipeline when needed. </blockquote></details> </blockquote></details> <sub>✏️ Tip: You can disable this entire section by setting `review_details` to `false` in your review settings.</sub> </details> <!-- This is an auto-generated comment by CodeRabbit for review status -->
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py
Show resolved
Hide resolved
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py
Show resolved
Hide resolved
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Fix all issues with AI agents
In `@livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py`:
- Around line 102-103: The code mutates instance state by assigning
self._opts.language = language when a per-call language override is provided;
instead compute a local effective_language variable (e.g., effective_language =
language if is_given(language) else self._opts.language) inside the method and
pass that to _transcription_to_speech_event (update
_transcription_to_speech_event signature to accept the effective_language) so
the instance's self._opts remains unchanged and per-call overrides only affect
the current call.
🧹 Nitpick comments (1)
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py (1)
193-200: Consider explicitly handlingasyncio.CancelledErrorfor clarity.While
asyncio.CancelledErroris aBaseExceptionsubclass (notException) in Python 3.8+, explicitly re-raising it improves code clarity and future-proofs against potential changes:♻️ Optional improvement
+import asyncio import base64 import uuid from dataclasses import dataclassexcept openai.APITimeoutError: raise APIConnectionError() from None + except asyncio.CancelledError: + raise except openai.APIStatusError as e: logger.error(f"TTS API error: {e}") raise APIConnectionError() from e
📜 Review details
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (5)
livekit-plugins/livekit-plugins-liquidai/README.mdlivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/py.typedlivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.pylivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.pylivekit-plugins/livekit-plugins-liquidai/pyproject.toml
🚧 Files skipped from review as they are similar to previous changes (2)
- livekit-plugins/livekit-plugins-liquidai/README.md
- livekit-plugins/livekit-plugins-liquidai/pyproject.toml
🧰 Additional context used
📓 Path-based instructions (1)
**/*.py
📄 CodeRabbit inference engine (AGENTS.md)
**/*.py: Format code with ruff
Run ruff linter and auto-fix issues
Run mypy type checker in strict mode
Maintain line length of 100 characters maximum
Ensure Python 3.9+ compatibility
Use Google-style docstrings
Files:
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.pylivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py
🧠 Learnings (3)
📓 Common learnings
Learnt from: CR
Repo: livekit/agents PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-01-16T07:44:56.353Z
Learning: Follow the Plugin System pattern where plugins in livekit-plugins/ are separate packages registered via the Plugin base class
📚 Learning: 2026-01-16T07:44:56.353Z
Learnt from: CR
Repo: livekit/agents PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-01-16T07:44:56.353Z
Learning: Implement Model Interface Pattern for STT, TTS, LLM, and Realtime models with provider-agnostic interfaces, fallback adapters for resilience, and stream adapters for different streaming patterns
Applied to files:
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py
📚 Learning: 2026-01-18T01:09:01.847Z
Learnt from: davidzhao
Repo: livekit/agents PR: 4548
File: livekit-plugins/livekit-plugins-openai/livekit/plugins/openai/responses/llm.py:77-81
Timestamp: 2026-01-18T01:09:01.847Z
Learning: In the OpenAI responses LLM (`livekit-plugins-openai/livekit/plugins/openai/responses/llm.py`), reasoning effort defaults are intentionally set lower than OpenAI's API defaults for voice interactions: "none" for gpt-5.1/gpt-5.2 and "minimal" for other reasoning-capable models like gpt-5, to avoid enabling reasoning by default in voice contexts.
Applied to files:
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.pylivekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py
🧬 Code graph analysis (1)
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py (4)
livekit-agents/livekit/agents/_exceptions.py (2)
APIConnectionError(84-88)APITimeoutError(91-95)livekit-agents/livekit/agents/types.py (1)
APIConnectOptions(54-88)livekit-agents/livekit/agents/utils/misc.py (1)
is_given(25-26)livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py (4)
model(76-77)provider(80-81)update_options(83-92)aclose(160-161)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (3)
- GitHub Check: type-check (3.9)
- GitHub Check: type-check (3.13)
- GitHub Check: unit-tests
🔇 Additional comments (11)
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py (6)
1-23: LGTM!Imports are well-organized and constants are appropriately defined for a local LiquidAI server setup.
25-73: LGTM!The initialization is well-structured with appropriate timeout configurations for audio processing. The use of
NOT_GIVENpattern for optional parameters aligns with the codebase conventions.
75-81: LGTM!Properties are correctly implemented.
83-92: LGTM!The
update_optionsmethod follows the established pattern for runtime configuration updates.
154-158: LGTM!The helper method correctly constructs the speech event.
160-161: LGTM!Proper cleanup of the async client.
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/tts.py (5)
1-37: LGTM!Imports and constants are well-defined. The audio parameters (24kHz sample rate, mono channel) are appropriate for TTS output.
39-81: LGTM!Initialization is consistent with the STT implementation, with appropriate TTS-specific capabilities and audio parameters.
102-117: LGTM!The
synthesizefactory method andaclosecleanup are correctly implemented.
120-134: LGTM!
ChunkedStreaminitialization properly stores the required dependencies for streaming.
136-191: LGTM!The streaming implementation correctly processes audio chunks, converting float32 PCM to int16 format. The defensive checks for
chunk.choicesanddeltahandle edge cases appropriately.
✏️ Tip: You can disable this entire section by setting review_details to false in your review settings.
livekit-plugins/livekit-plugins-liquidai/livekit/plugins/liquidai/stt.py
Outdated
Show resolved
Hide resolved
2 participants
Added LiquidAI LFM2.5-Audio plugin for LiveKit Agents
Summary by CodeRabbit
New Features
Documentation
Chores
✏️ Tip: You can customize this high-level summary in your review settings.