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
Transcribe pre-recorded audio or live streams in a few lines of Python. One API returns accurate text with timestamps, speaker labels, and language detection—at any scale. Start free, no credit card.
Walk me through the last time the export feature failed for you.
It was during the Meridian Labs onboarding—the file just never arrived.
Got it. And how often does that happen in a typical week?
Send a local file or URL—MP3, WAV, MP4, and more—and get back formatted text with word-level timestamps and speaker labels.
Stream mic or telephony audio over a WebSocket and transcripts come back within a few hundred milliseconds.
pip install assemblyai covers both—the SDK handles upload, submission, and polling, so there's no queue plumbing to write.
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.”
Open-source libraries like SpeechRecognition, Vosk, and Whisper are a fine way to prototype—if you're up for managing models, GPUs, and accuracy tradeoffs yourself. An API moves that work onto managed infrastructure: you send audio, and accurate transcripts come back with speakers, timestamps, and 99+ languages—whether you're processing one file or a million.
Start building freeNo credit card required
Word error rate is the share of words the model gets wrong against a human reference—the standard measure of transcription accuracy on pre-recorded audio.
Pre-recorded word error rate on English audio.
*Lower is better*Source: AssemblyAI published benchmarks — assemblyai.com/benchmarks.
