Picovoice Leopard
On-device offline speech-to-text from Picovoice — file-based, no cloud.
On-device offline speech-to-text from Picovoice — file-based, no cloud.
Best for offline batch transcription with diarization on devices that can't ship audio to the cloud. Pricing: Free tier + per-user/Enterprise plans.
What it is
Picovoice Leopard is the offline batch counterpart to Cheetah, taking complete audio files and producing transcripts with word timestamps and speaker diarization, all running on-device. It targets compliance-sensitive customers (healthcare, legal) who want a local-only option. Same access-key licensing as the rest of the Picovoice stack. Best fit: offline batch transcription with diarization on devices that can't ship audio to the cloud. Caveats: smaller language coverage than whisper; requires picovoice access key. Pricing as listed: Free tier + per-user/Enterprise plans. Feature flags from vendor docs: speaker diarization, word-level timestamps. Directory tags: embedded, on-device. Last vendor-page check: 2026-05-12.
Watch out for: Smaller language coverage than Whisper; requires Picovoice access key.
Install / use
pip install pvleopardFeatures
| Speaker diarization | Yes |
| Word-level timestamps | Yes |
| Streaming / real-time | No |
| Languages supported | 7 |
| HIPAA eligible | No |
Picovoice Leopard vs Whipscribe
| Feature | Picovoice Leopard | Whipscribe |
|---|---|---|
| Category | Products | Transcription APIs |
| Pricing | Free tier + per-user/Enterprise plans | free beta |
| Speaker diarization | Yes | Yes |
| Word timestamps | Yes | Yes |
| Streaming | No | No |
| Languages | 7 | 99 |
| Platforms | Edge, SDK, Linux, Windows, macOS | Web, API, MCP |
Alternatives to Picovoice Leopard
Whipscribe is a managed faster-whisper + whisperX service. If you want transcripts without running infrastructure, paste a URL or drop a file in the form below — you'll have a transcript in seconds.