Speechmatics
Speech recognition engine with strong accent coverage and on-premise deployment.
why this verdict
Keep it for the workflow, not the AI
Whisper covers the headline feature free, and the workflow around it is why you would still open this.
- A named launch, not a vibe named, unlinked
- Whisper being free and open source, collapsing the price of speech recognition. OpenAI, October 1, 2025 — no announcement link recorded yet.
- How much of the job it covers headline only editorial call
- The headline feature, but not the workflow, data or integrations around it.
- Is there a free way to do it? yes
- 1 of 3 listed replacements have a usable free tier: Whisper.
- What the call is worth nothing to cancel
- No paid entry tier tracked, so there is no subscription to cancel.
- Threatened by
- OpenAI
- Since
- October 1, 2025
- List price
- free
- Per year
- —
The backstory
Speechmatics built genuinely excellent speech recognition with unusually good accent and dialect handling, sold to broadcasters and enterprises that need transcription on their own infrastructure. Then OpenAI released Whisper for free, and the floor fell out of per-minute transcription pricing everywhere. What survives is the part Whisper does badly: real-time streaming at scale, on-premise deployment, and the accuracy guarantees a broadcaster's captions contract actually requires.
Escape hatches
Fast streaming recognition with competitive pricing
deepgram.com open_in_newStrong developer experience and audio intelligence
assemblyai.com open_in_newFree open-source model you can run yourself
github.com open_in_new1 of 3 replacements have a usable free tier.