Best Open-Source Speech-to-speech options.
Candidates with open-source, open-weight, downloadable pretrained, or self-hosted signals from source metadata. Automated source-backed radar for S2S, voice-to-voice, live audio translation, and real-time voice model candidates. Rankings use existing source-backed radar data and do not invent prices, benchmark scores, release dates, or capabilities.
LiveKit Agents
Highest automated fit score using source-backed task evidence and radar signals.
LiveKit Agents
Strong candidate when the source metadata indicates API or hosted-service availability.
LiveKit Agents
Strong candidate when open weights, GitHub, permissive license, or self-hosting signals are visible.
| # | Candidate | Type | Access | Source | Fit | Updated | Why it matched | Evidence |
|---|---|---|---|---|---|---|---|---|
| 01 | LiveKit AgentsLiveKit framework | service | Paid API | Official service page | Score71 | Jun 26, 2026 | Matched speech-to-speech, speech to speech; 2 source links; official service signal; access model: Paid API | docs.livekit.io |
Track Speech-to-speech changes
The model, API, pricing and tooling changes that matter, in one morning email with links to the original sources. Free, unsubscribe in one click.
Prefer topic alerts or a custom RSS feed?