- Home
- AI transcription tools
- IBM Watson Speech to Text
On the recordRead 9 Oct 202622 facts
IBM Watson Speech to Text
IBM Watson Speech to Text · Armonk, New York, United States · founded 1911
Plans and prices
Plus
$0.02usage-based
Unlimited minutes · 100 concurrent transcription requests · custom language and acoustic models
Read 9 Oct 2026 · cloud.ibm.com
Lite
Free
500 minutes per month · no customization · services deleted after 30 days of inactivity
Read 9 Oct 2026 · cloud.ibm.com
Premium
Priced on request
Unlimited minutes · unlimited concurrent transcriptions · data isolation and added security features · HIPAA support available in specified locations
Read 9 Oct 2026 · cloud.ibm.com
On the record
- API access
- Yescloud.ibm.com · read 22 Sep 2026
- Call analysis
- IBM describes using Speech to Text transcriptions with Watson Natural Language Understanding to analyze customer calls.cloud.ibm.com · read 9 Oct 2026
- Customization
- Paid plans can customize language and acoustic models, and language models can use grammars to restrict recognized phrases.cloud.ibm.com · read 9 Oct 2026
- Data deletion
- Customers can label data with a customer ID and use an API method to delete data associated with that ID.cloud.ibm.com · read 9 Oct 2026
- Export formats
- JSONcloud.ibm.com · read 22 Sep 2026
- Input limitation
- The service accepts audio data only; IBM says audio must be separated from video before transcription.cloud.ibm.com · read 9 Oct 2026
- Integrations
- IBM documents a pre-built phone assistant integration combining watsonx Assistant, Speech to Text, and Watson Text to Speech.cloud.ibm.com · read 9 Oct 2026
- Intended users
- IBM identifies customer self-service, agent assistance, and call analytics as use cases, and describes Plus as intended for small businesses and development or testing by larger enterprises.ibm.com · read 9 Oct 2026
- Interim results
- The product page says applications can use transcription as it is generated before final results are complete.ibm.com · read 9 Oct 2026
- Languages supported
- 14cloud.ibm.com · read 22 Sep 2026
- Languages and audio
- IBM lists support for multiple languages and formats including WAV, FLAC, MP3, Ogg, WebM, and PCM.cloud.ibm.com · read 9 Oct 2026
- Request limits
- Synchronous HTTP and WebSocket requests accept up to 100 MB of audio, while asynchronous HTTP accepts up to 1 GB.cloud.ibm.com · read 9 Oct 2026
- Security
- Premium offers data isolation and features including Key Protect, service endpoints, mutual authentication, and HIPAA readiness.cloud.ibm.com · read 9 Oct 2026
- Speaker diarization
- IBM’s product page says the service can recognize who said what in a multi-participant voice exchange.ibm.com · read 9 Oct 2026
- Speaker identification
- Yescloud.ibm.com · read 22 Sep 2026
- Streaming and batch
- It supports synchronous and asynchronous HTTP interfaces and a full-duplex WebSocket interface for low-latency communication.cloud.ibm.com · read 9 Oct 2026
- Support availability
- Premium instances require contacting an IBM sales representative before provisioning.cloud.ibm.com · read 9 Oct 2026
- Timestamp support
- Yescloud.ibm.com · read 22 Sep 2026