Polish speech recognition for contact centers
Turn phone calls into accurate, punctuated transcripts with word timestamps and per-speaker channels, through an OpenAI-compatible API. Built by VoiceLab, processed in Poland.
Everything a contact center needs from ASR
One model tuned for real phone calls, delivered as a simple, predictable API.
Built for telephony
Polish acoustic model for 8 kHz call-center audio: punctuated, formatted text with word timestamps and confidence scores.
Who said what
Stereo and multi-channel recordings are transcribed per channel, so agent and customer stay apart, even when they talk over each other.
OpenAI-compatible API
Use the OpenAI SDK or plain HTTP. JSON, text, verbose JSON with segments and words, SRT and VTT.
Processed in Poland
Recognition runs on VoiceLab servers in the EU. Recordings and transcripts are not stored and never used for training.
Long recordings
Up to 60 minutes and 256 MB per request: wav, mp3, flac, ogg/opus, m4a, webm, amr.
Prepaid and transparent
Billed per second of audio. Your balance, usage and every request are visible in the dashboard.
Start in three steps
No contract and no sales call needed to try it.
Sign in with Google
Your account is created at once with 1 hour of free audio.
Create an API key
Keys are shown once and can be revoked at any time.
Send audio
POST a recording, get the transcript back in seconds.
# transcribe a call curl https://asr.voicelab.ai/api/v1/audio/transcriptions \ -H "Authorization: Bearer $VOICELAB_API_KEY" \ -F file=@call.wav -F model=voicelab/moonlight-lvcsr
Try it on your own recordings
1 hour of audio free to start. Need volume pricing? Write to voicelab@voicelab.ai.