LiveKit Agents builds voice agents that join LiveKit rooms, mixing STT, LLM, TTS or realtime APIs via plugins. Code is Apache-2.0; its turn-detection models use the LiveKit Model License (only with LiveKit Agents). The README example uses LiveKit Inference (usage-billed, in LiveKit Cloud) with Deepgram, Cartesia and Gemma. LiveKit server can be self-hosted and docs cover a local Ollama LLM; no local STT/TTS plugin was found.
Cite this pageRepoQuarry. “LiveKit Agents.” Checked 16 Oct 2026. https://repoquarry.dev/projects/livekit-agents
How much of ElevenLabs does it replace?
1 of 4 jobs. Framework only; STT and TTS come from plugged-in providers, which in the documented plugins are hosted services.
No:Speech generationNot covered
No:Voice cloningNot covered
No:Speech recognitionNot covered
Partial:Voice agentsPartial replacement
For speech generation, voice cloning, speech recognition, see other ElevenLabs alternatives. You may also need to build an interface, an API server and hosting yourself.
How much of Vapi does it replace?
2 of 2 jobs. Framework plus LiveKit SIP telephony stack; STT/TTS come from plugged-in providers.
Runs on my machine?Confirmed:Yes; console mode runs locally, and the full stack including LiveKit server can be self-hosted; LiveKit Cloud is the hosted optionFully open-source, allowing you to run the entire stack on your own servers, including LiveKit server, one of the most widely used WebRTC media servers.github.com · agents#features ↗1d ago
Code licenceConfirmed:Apache-2.0The Agents framework is licensed under Apache-2.0.github.com · LICENSE ↗1d ago
Model weights licencePartial:LiveKit turn-detection models are under the LiveKit Model License: royalty-free, but only usable with LiveKit Agents and not to train other modelsyou may use these LiveKit models freely but can only use them together with the LiveKit Agents framework. You cannot use the LiveKit models on a standalone basis or with any other frameworks.github.com · MODEL_LICENSE ↗1d ago
Commercial usePartial:Apache-2.0 framework; bundled turn-detector models carry LiveKit Model License restrictions; plugged-in providers have their own termsThe LiveKit turn detection models are licensed under the LiveKit Model License.github.com · agents#license ↗1d ago
Paid services needed?No:README example uses LiveKit Inference (part of LiveKit Cloud, billed by usage) or provider keys for Deepgram, OpenAI and Cartesia. LiveKit server can be self-hosted and docs cover a local Ollama LLM, but no local STT or TTS plugin is documented in the sources checkedLiveKit Inference billing is based on usage.docs.livekit.io · inference.md ↗1d ago
HardwarePartial:Framework requirements not stated; the turn-detector model runs on CPU and needs under 500 MB RAMThe model requires <500MB of RAM and runs within a shared inference server, supporting multiple concurrent sessions.github.com · livekit-plugins-turn-detector ↗1d ago
Operating systemsPartial:Official voice-agent quickstart includes macOS, Linux and Windows setup paths, with Python/Node agent run instructions; selected model/transport plugins can add platform requirementsdocs.livekit.io · voice-ai ↗today
Setup effortConfirmed:Easy; pip install livekit-agents with provider extras, then set LiveKit URL and API key/secretTo install the core Agents library, along with plugins for popular model providersgithub.com · agents#installation ↗1d ago
Your choice on cookies
We use essential storage to keep the site secure and remember this choice. With your permission we'd also count visits to see which pages help people, and remember your filters. We never use advertising or cross-site tracking cookies, and sponsors never get your data. Cookie policy