livekit-plugins-inworld

What it is and what it does

This package is a LiveKit Agents plugin that wraps Inworld's text-to-speech and speech-to-text APIs, allowing you to build voice agents that synthesize speech and transcribe user input using Inworld's models. It handles the integration plumbing so you can drop Inworld TTS and STT components into a LiveKit AgentSession alongside an LLM and other agent logic.

The TTS side supports multiple voice IDs, encoding formats (LINEAR16, MP3, OGG_OPUS, ALAW, MULAW, FLAC), sample rates, and real-time streaming via WebSocket with buffering controls for lower-latency synthesis. The STT side provides streaming speech-to-text with optional voice profile detection. Both are designed to work within LiveKit's real-time agent framework, so you configure them as session components and let the framework handle audio routing.

Use it for:

• Build a voice agent that listens to user speech, transcribes it with Inworld STT, processes it through an LLM, and speaks back using Inworld TTS.
• Stream text incrementally to Inworld TTS as it's generated by an LLM, reducing latency in conversational interactions.
• Integrate Inworld's voice synthesis into an existing LiveKit agent to replace or supplement other TTS providers.
• Deploy a multi-modal agent that combines LiveKit's video/audio infrastructure with Inworld's speech models for real-time conversations.
• Use voice profile detection with Inworld STT to identify speaker characteristics during transcription in multi-participant sessions.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Integrates Inworld's text-to-speech and speech-to-text APIs into LiveKit voice agents, enabling real-time voice synthesis and transcription for conversational AI applications.

Yes, if you are building a LiveKit voice agent and want to use Inworld's TTS and STT APIs. The package is actively maintained, has no known vulnerabilities, installs cleanly with minimal dependencies, and the Apache-2.0 license is permissive. The main prerequisite is an Inworld API key and willingness to use their services; evaluate Inworld's pricing and model quality for your use case before committing.

Install

livekit-plugins-inworld on PyPI

pip install livekit-plugins-inworld

uv add livekit-plugins-inworld

poetry add livekit-plugins-inworld

Installing livekit-plugins-inworld

Before you install

Low friction: pure Python wheel with a single runtime dependency (livekit-agents). Released 1 day ago with active maintenance; the parent repository has 13004 stars and current commit history.

License in practice

Apache-2.0 permissive license allows commercial and private use with minimal restrictions—suitable for most production deployments.

Quickstart

pip install livekit-plugins-inworld

from livekit.plugins import inworld

tts = inworld.TTS(voice="Hades")
stt = inworld.STT(model="inworld/inworld-stt-1")

Requires INWORLD_API_KEY environment variable set in .env file; obtain from https://platform.inworld.ai/login. Requires Python >=3.10.0.

Verify before relying

• Whether Inworld API pricing or rate limits affect typical agent workloads.
• Supported voice IDs and whether custom cloned voices are available in all Inworld plans.
• Latency characteristics of WebSocket streaming mode under typical network conditions.

Package facts

Evidence: livekit_plugins_inworld-1.6.10-py3-none-any.whl

Keywords: ai, audio, inworld, livekit, realtime, speech-to-text, stt, tts, video, voice

Tags

More Artificial Intelligence packages

Further reading