Compare realtime APIs
Pick up to four APIs from any category and see them side by side.
| Picovoice Cheetah (on-device streaming STT) Picovoice | Amazon Transcribe Streaming (incl. Medical and Call Analytics) AWS | Azure AI Speech real-time speech to text Microsoft | |
|---|---|---|---|
| Category | Speech-to-text | Speech-to-text | Speech-to-text |
| Status | GA | GA | GA |
| Est. per minute | $0.02 | $0.01 - 0.075 | $0.0067 - 0.025 |
| How that was worked out | Foundation Plan if fully used: $6,000 / (25,000 min x 12) = $0.02/min. Unused minutes raise the effective rate. | $0.01/min standard streaming up to $0.075/min Medical. Add PII redaction or CLM as needed. | $0.40/hr commitment-tier overage up to $1.20/hr custom + $0.30/hr add-on. Pay-as-you-go standard is $0.0167/min. |
| Pricing model | free | per-minute | per-hour |
| Free tier | Non-commercial free tier; commercial free trial grants Foundation Plan rights. | 60 minutes per month for 12 months from first request (excludes PII redaction). | F0: 5 audio hours per month shared between standard and custom real-time; 1 concurrent request. |
| Connects by | On-device SDK (no network audio) | HTTP/2, WebSocket | WebSocket (via Speech SDK), REST for short audio |
| Audio in | 16-bit PCM frames of Cheetah's frame_length (16 kHz mono in Picovoice SDKs; not restated on the doc page read). | PCM signed 16-bit little-endian (not WAV), FLAC, or Opus in Ogg. 16 kHz recommended; 8 kHz telephony accepted. Chunks of 50-200 ms recommended; send zero-byte silence rather than pausing. | Speech SDK handles microphone, files and push/pull streams; default 16 kHz 16-bit mono PCM; compressed formats via GStreamer. |
| Audio out | n/a | n/a | n/a |
| Languages | English, French, German, Italian, Japanese, Korean, Spanish, Portuguese; more for Enterprise. | Many streaming languages (count not re-verified); Medical is US English. | 100+ locales (see language support page; count not re-verified). |
| Latency (vendor claim) | Runs locally; no network round trip. No numeric claim captured. | No vendor figure captured; latency depends on chunk size per docs. | No vendor figure captured. |
| Key limits |
|
|
|
| High-severity warnings |
|
|
|
| Compliance | Audio processed on device (privacy by design). | Amazon Transcribe and Transcribe Medical are HIPAA-eligible under the AWS BAA per AWS (not re-verified here). AI services opt-out policy controls data use for training. | Azure compliance scope (HIPAA BAA, SOC, ISO) applies to Azure AI Speech per Microsoft; connected/disconnected containers available. Not re-verified. |
| Self-hostable | Yes | No | Yes |
| Last checked | 2026-10-10 | 2026-10-10 | 2026-10-10 |
| Key numbers and features | |||
| $/hour | $1.2 | $0.6 | $1 |
| Bills silence | - | Yes | - |
| Free tier | Yes | Yes | Yes |
| Free credit $ | - | - | - |
| Live speakers | - | Yes | Yes |
| Turn detect | Yes | - | - |
| Keyterms | Yes | Yes | Yes |
| PII redact | - | Yes | - |
| Partials | Yes | Yes | Yes |
| Mixed langs | - | - | Yes |
| 8 kHz phone | - | No | - |
| Latency ms | - | - | - |
| Languages | 8 | 77 | - |
| Max session min | - | - | - |
| Concurrency | - | 25 | 100 |
| WebRTC | No | No | No |
| WebSocket | No | Yes | Yes |
| gRPC | No | No | No |
| HIPAA | - | Yes | Yes |
| SOC 2 | - | - | Yes |
| EU data | - | Yes | Yes |
| Self-host | Yes | No | Yes |
| Open weights | No | No | No |
| High warnings | 1 | 1 | 1 |