customers
All customer stories
Top Voice AI companies are building with Assembly.
resources
Latest Release
Universal 3.5 Pro
Our new flagship speech-to-text model handles real-world audio the way it actually happens. Available for real-time and pre-recorded audio.
resources
One WebSocket gives you speech recognition, turn detection, interruption handling, and tool calling, built on AssemblyAI's real-time Speech Understanding—so you can talk to a working agent today.
Hi, I need to move my appointment to next Thursday.
Sure—I can see your booking. Thursday at 2pm works. Should I confirm it?
Yes, and send the confirmation to dana@acme.io.
Speech recognition, turn detection, LLM reasoning, and voice over a single connection, so there's no STT, LLM, or TTS stack to wire together.
Built-in turn detection and barge-in handling let users interrupt and be understood the way real conversations work.
Hand the agent HTTP tools and it can look up, book, and act mid-conversation without extra orchestration code.
36%
improvement in close rate
“The new Universal-3.5 Pro speech model from AssemblyAI is best so far in terms of accuracy, latency, and language switching.”
80%
increase in customer satisfaction
“Assembly has saved us countless hours managing models, and provided exceptional accuracy.”
36%
improvement in close rate
“The new Universal-3.5 Pro speech model from AssemblyAI is best so far in terms of accuracy, latency, and language switching.”
80%
increase in customer satisfaction
“Assembly has saved us countless hours managing models, and provided exceptional accuracy.”
Create an agent with one API call, drop it into a small web page, and you're in a live conversation in minutes—echo cancellation is handled for you. Or try a live agent in the playground first.
Start building freeNo credit card required
A voice agent is only as good as the words it hears. Streaming word error rate measures real-time transcription accuracy against a human reference—errors here compound through every downstream step.
Streaming word error rate on English audio.
*Lower is better*Source: AssemblyAI published benchmarks — assemblyai.com/benchmarks.
