Speaker Diarization API
pyannote 3.1, standalone — no transcription bundling required
Identify who spoke when. Just speaker labels — not text. Powered by pyannote 3.1 (research SOTA on AMI), at $0.010/min — 4.5× cheaper than pyannote.ai, 6.5× cheaper than AssemblyAI.
How we compare
Standalone diarization is rare — most providers bundle it with transcription. Brainiall exposes pyannote 3.1 directly so you only pay for speaker labels. v1.0 calibration vs published competitor pricing and metrics; v1.1 will replace with direct DER measurement on a 200-clip benchmark.
| Provider | Quality | Price/minute | vs market avg | Position |
|---|---|---|---|---|
| pyannote.ai | 9.0/10 | $0.045 | 131% | — |
| AssemblyAI | 8.5/10 | $0.065 | 190% | — |
| Rev.ai | 8.7/10 | $0.020 | 58% | — |
| AWS Transcribe | 8.0/10 | $0.024 | 70% | — |
| Deepgram Nova-3 + diarize | 8.6/10 | $0.018 | 51% | — |
| Brainiall FAST | 9.0/10 | $0.010(71% cheaper) | 29% | Parity |
Pricing rule: 90% off when inferior · 80% off at parity · 50% off when superior. Position determined by objective benchmark, refreshed quarterly.
Pricing
Discount derived from quality position vs the closest competitor. 90% off when inferior, 80% off at parity, 50% off when superior.
Free
$0/mo
30 min/month · fast tier · forever free
Starter
$19/mo
1,000 min/month · turns + RTTM + LAB output · all formats
Pro
$99/mo
10,000 min/month · priority queue · 99.5% SLA
Business
$299/mo
50,000 min/month · dedicated capacity · email + Slack
PAYG: $0.010/min (pyannote 3.1). HD tier (pyannote + ECAPA-TDNN re-clustering for ~3pp DER improvement on overlapping audio) is on the v1.1 roadmap — not yet available.
One endpoint, three output formats
# Standard JSON turns
POST https://api.brainiall.com/v1/audio/diarize/base64
{"audio": "<base64 wav>", "tier": "fast"}
# Industry-standard RTTM (Rich Transcription Time Marked)
POST https://api.brainiall.com/v1/audio/diarize/base64
{"audio": "<base64>", "output_format": "rttm"}
# With speaker count hint (improves accuracy on known-cardinality calls)
POST https://api.brainiall.com/v1/audio/diarize/base64
{"audio": "<base64>", "num_speakers": 2}Why standalone diarization?
- Already have transcripts: Whisper output + speaker labels is the most cost-efficient pipeline.
- Privacy-regulated audio: get speaker analytics without sending text-bearing audio to a transcription provider.
- Multi-language workflows: speaker change detection is language-agnostic; works on any language pyannote supports.
- Cost: paying transcription rates ($0.024-0.65/min) for diarization is overkill when standalone is $0.010/min.
Press kit & resources
Everything reviewers, integrators and procurement teams typically ask for.
One-page datasheet
Pricing, KPIs and a copy-pasteable curl snippet on one page. Ideal for buyer review.
Download PDFTry it with our sample
60-second 3-speaker recording — feed it through the API and compare the output against your own input.
Download sampleAPI reference
OpenAPI spec, request/response shapes, error codes, rate limits and quota model.
Read docs →More specialty APIs
Same single API key, same usage-based pricing, different problem solved.
