LiveKit Plugins Hume

1.5.8 verified Sat May 09 auth: no python

Hume AI TTS plugin for LiveKit Agents, enabling text-to-speech synthesis using Hume's emotional intelligence models. Current version 1.5.8, requires Python >=3.10. Part of the LiveKit Agents ecosystem with monthly releases.

pip install livekit-plugins-hume

Common errors

error ModuleNotFoundError: No module named 'livekit.plugins.hume' ↓

cause Package not installed or installed incorrectly.

fix

Run pip install livekit-plugins-hume

error TypeError: TTS() got an unexpected keyword argument 'preemptive_generation' ↓

cause `preemptive_generation` removed in v1.5.0.

fix

Remove argument; use agent PreemptiveGenerationOptions.

error livekit.plugins.hume.TTS requires 'HUME_API_KEY' environment variable ↓

cause API key not provided and not set in environment.

fix

Set HUME_API_KEY environment variable or pass api_key parameter.

error AttributeError: module 'livekit.plugins.hume' has no attribute 'tts' ↓

cause Old import path lowercase 'tts' no longer exists.

fix

Use from livekit.plugins.hume import TTS (uppercase).

Warnings

gotcha Deprecation notice: LiveKit Agents v1.5 removed `preemptive_generation: bool` from TTS classes. Use `agent/voice_pipeline_agent.PreemptiveGenerationOptions` instead. ↓

fix Migrate to `PreemptiveGenerationOptions` (see LiveKit agents v1.5 changelog).

gotcha The `api_key` parameter is no longer read from `HUME_API_KEY` environment variable automatically. You must pass it explicitly. ↓

fix Set `api_key=os.environ.get('HUME_API_KEY')` when constructing TTS.

deprecated Old import path `livekit.plugins.hume.tts` (lowercase) is deprecated. Use `livekit.plugins.hume.TTS` (uppercase). ↓

fix Use `from livekit.plugins.hume import TTS`.

Imports

TTS
```
from livekit.plugins.hume import TTS
```
No common wrong import pattern

Quickstart

Initialize Hume TTS with API key and use with AgentSession.

import os
from livekit import agents
from livekit.plugins.hume import TTS

async def main():
    tts = TTS(
        api_key=os.environ.get('HUME_API_KEY', ''),
    )
    # Use with AgentSession or VoicePipelineAgent
    session = agents.VoicePipelineAgent(
        stt=None,
        llm=None,
        tts=tts,
    )
    await session.start()