Speech-to-speech model and API for voice agents that skips the transcription step.
Ultravox is voice-agent infrastructure from Fixie.ai. Its model consumes audio directly instead of running speech through a transcriber first, which removes a hop from the response path and preserves tone and interruption handling that a text pipeline flattens. Developers get a REST API, web and mobile SDKs, SIP for telephone calls, RAG corpora for grounding answers and custom voices; the underlying open-weight models are published on Hugging Face, so the core can be self-hosted rather than rented. Billing is per minute of conversation with text-to-speech included, plus separate token rates for the text threads feature. Fixie.ai Inc. hosts the service in the United States.
Pay as you go has no monthly fee and charges $0.05 per minute of conversation including text-to-speech, with a free start capped at 5 concurrent calls. Pro is $100 a month and adds 5 custom voices and 20 RAG corpora. SIP telephony is $0.005 per minute; threads are $2 per million input tokens and $15 per million output tokens.
Use tool βOne of the few voice stacks where the model itself is the product rather than glue over someone else's transcriber, and the open weights on Hugging Face mean you are not locked into the hosted service. Skipping the speech-to-text hop is the whole point β latency and prosody both benefit. It is a builder's tool, not a no-code agent designer, so budget engineering time as well as per-minute cost.
Watch out: The old Fixie.ai domain now redirects here β Fixie's earlier agent products are gone and Ultravox is what the company sells.
Side-by-side fact sheets against the competitors our research names β pricing, platforms, API and both verdicts on one page.
No reviews yet β be the first to review Ultravox.ai.
Reviews are tied to your EffectHub account β one review per tool, so the rating for Ultravox.ai reflects real users.
No questions yet β ask the first one about Ultravox.ai.
Questions about Ultravox.ai are tied to your EffectHub account, so answers can reach you and the section stays free of spam.