Multimedia Tts (lexigram-multimedia-tts)

Text-to-speech generation for the Lexigram Framework — local and API-based backends (local-http, elevenlabs, openai, chatterbox, kokoro, f5-tts, piper).

Overview

lexigram-multimedia-tts synthesizes speech from text. The default backend calls a local HTTP reference server (http://localhost:5002) so the package works out of the box with no API keys; hosted backends (ElevenLabs, OpenAI) and in-process reference servers (Chatterbox, Kokoro, F5-TTS, Piper) are selectable via config.

Full documentation: docs.lexigram.dev

Install

uv add lexigram-multimedia-tts
# Optional extras
uv add "lexigram-multimedia-tts[elevenlabs]"        # ElevenLabs API
uv add "lexigram-multimedia-tts[openai]"            # OpenAI TTS API
uv add "lexigram-multimedia-tts[chatterbox-server]" # local Chatterbox server (torch)
uv add "lexigram-multimedia-tts[kokoro-server]"     # local Kokoro server
uv add "lexigram-multimedia-tts[f5-tts-server]"     # local F5-TTS server (torch)
uv add "lexigram-multimedia-tts[piper-server]"      # local Piper server

Quick Start

from lexigram import Application
from lexigram.di.module import Module, module
from lexigram.multimedia.tts import AudioTTSModule
from lexigram.contracts.multimedia import TTSProvider, TTSRequest


@module(imports=[AudioTTSModule.configure()])
class AppModule(Module):
    pass


async def main() -> None:
    async with Application.boot(modules=[AppModule]) as app:
        tts = await app.container.resolve(TTSProvider)
        result = await tts.generate(TTSRequest(text="Hello from Lexigram", voice="alloy"))
        if result.is_ok():
            asset = result.unwrap()  # MediaAsset — audio bytes or URI


if __name__ == "__main__":
    import asyncio

    asyncio.run(main())

Configuration

Zero-config usage: Call AudioTTSModule.configure() with no arguments to use the local-http backend at http://localhost:5002.

Option 1 — YAML file

multimedia:
  tts:
    backend: "elevenlabs"
    elevenlabs_voice_id: "21m00Tcm4TlvDq8ikWAM"

Option 2 — Profiles + Environment Variables

export LEX_PROFILE=production
export LEX_MULTIMEDIA__TTS__BACKEND=openai
export LEX_MULTIMEDIA__TTS__OPENAI_VOICE=echo

Option 3 — Python

from lexigram.multimedia.tts import AudioTTSModule
from lexigram.multimedia.tts.config import TTSConfig

AudioTTSModule.configure(
    config=TTSConfig(backend="elevenlabs", elevenlabs_voice_id="21m00Tcm4TlvDq8ikWAM")
)

Config reference

Field	Default	Env var	Description
`backend`	`"local-http"`	`LEX_MULTIMEDIA__TTS__BACKEND`	`local-http`, `elevenlabs`, `openai`, `chatterbox`, `kokoro`, `f5-tts`, `piper`
`local_http_base_url`	`"http://localhost:5002"`	`LEX_MULTIMEDIA__TTS__LOCAL_HTTP_BASE_URL`	Local reference server URL
`elevenlabs_voice_id`	`None`	`LEX_MULTIMEDIA__TTS__ELEVENLABS_VOICE_ID`	ElevenLabs voice ID (required for `elevenlabs`)
`elevenlabs_api_key_secret_name`	`"elevenlabs_api_key"`	`LEX_MULTIMEDIA__TTS__ELEVENLABS_API_KEY_SECRET_NAME`	Secret name for the ElevenLabs API key
`openai_api_key_secret_name`	`"openai_api_key"`	`LEX_MULTIMEDIA__TTS__OPENAI_API_KEY_SECRET_NAME`	Secret name for the OpenAI API key
`openai_voice`	`"alloy"`	`LEX_MULTIMEDIA__TTS__OPENAI_VOICE`	OpenAI voice (`alloy`, `echo`, `fable`, `onyx`, `nova`, `shimmer`)
`openai_model`	`"tts-1"`	`LEX_MULTIMEDIA__TTS__OPENAI_MODEL`	OpenAI TTS model
`openai_base_url`	`"https://api.openai.com"`	`LEX_MULTIMEDIA__TTS__OPENAI_BASE_URL`	OpenAI-compatible base URL
`chatterbox_base_url`	`"http://localhost:5100"`	`LEX_MULTIMEDIA__TTS__CHATTERBOX_BASE_URL`	Chatterbox server URL
`chatterbox_exaggeration`	`0.5`	`LEX_MULTIMEDIA__TTS__CHATTERBOX_EXAGGERATION`	Chatterbox exaggeration factor
`chatterbox_cfg_weight`	`0.5`	`LEX_MULTIMEDIA__TTS__CHATTERBOX_CFG_WEIGHT`	Chatterbox classifier-free guidance weight
`chatterbox_temperature`	`0.85`	`LEX_MULTIMEDIA__TTS__CHATTERBOX_TEMPERATURE`	Chatterbox sampling temperature
`kokoro_base_url`	`"http://localhost:5101"`	`LEX_MULTIMEDIA__TTS__KOKORO_BASE_URL`	Kokoro server URL
`kokoro_default_voice`	`"af_heart"`	`LEX_MULTIMEDIA__TTS__KOKORO_DEFAULT_VOICE`	Default Kokoro voice
`f5_tts_base_url`	`"http://localhost:5102"`	`LEX_MULTIMEDIA__TTS__F5_TTS_BASE_URL`	F5-TTS server URL
`piper_base_url`	`"http://localhost:5103"`	`LEX_MULTIMEDIA__TTS__PIPER_BASE_URL`	Piper server URL
`piper_default_voice`	`"en_US-lessac-medium"`	`LEX_MULTIMEDIA__TTS__PIPER_DEFAULT_VOICE`	Default Piper voice
`timeout`	`60.0`	`LEX_MULTIMEDIA__TTS__TIMEOUT`	Request timeout in seconds

Module Factory Methods

Method	Description
`AudioTTSModule.configure(config)`	Configure with explicit TTS config
`AudioTTSModule.stub()`	Real module pinned to the default `local-http` backend for tests

Key Features

Seven backends — local-http, elevenlabs, openai, chatterbox, kokoro, f5-tts, piper
Reference servers — lexigram-tts-*-serve console scripts run each local model server
Secret-managed API keys — provider keys resolved by name through the secrets backend
Result-based — generate() -> Result[MediaAsset, MultimediaError]; errors are domain values, not exceptions

Testing

from lexigram import Application
from lexigram.multimedia.tts import AudioTTSModule

async def test_boot():
    async with Application.boot(modules=[AudioTTSModule.stub()]) as app:
        assert app.container is not None

Key Source Files

File	What it contains
`src/lexigram/multimedia/tts/module.py`	`AudioTTSModule.configure()` and `.stub()`
`src/lexigram/multimedia/tts/config.py`	`TTSConfig`
`src/lexigram/multimedia/tts/di/provider.py`	`AudioTTSProvider` — registers `TTSProvider`, wires task handlers
`src/lexigram/multimedia/tts/providers/`	Backend implementations (`local_http`, `elevenlabs`, `openai`, `chatterbox`, `kokoro`, `f5_tts`, `piper`)
`src/lexigram/multimedia/tts/servers/`	Reference-server entry points (`lexigram-tts-*-serve`)
`src/lexigram/multimedia/tts/tasks.py`	Background generation task handlers
`src/lexigram/multimedia/tts/exceptions.py`	`TTSError` hierarchy