Modal (ASR endpoints)
Modal's serverless GPU platform commonly used to host Whisper / faster-whisper as an API.
Modal's serverless GPU platform commonly used to host Whisper / faster-whisper as an API.
Best for teams self-hosting Whisper / faster-whisper / WhisperX who want serverless GPU without managing K8s. Pricing: per-second GPU compute (see Modal pricing).
What it is
Modal Labs offers serverless GPU infrastructure used widely to host open-source ASR models (Whisper, faster-whisper, WhisperX, Pyannote diarization) behind a custom HTTP endpoint. You author Python code that loads a model and Modal handles scaling, cold-start, and billing per second. Not an ASR API itself, but a popular pick for teams that want the open-source stack with managed scaling. Pricing is per-second compute by GPU class. Best fit: teams self-hosting whisper / faster-whisper / whisperx who want serverless gpu without managing k8s. Caveats: not a turnkey asr api — you bring your own model code and pay for compute, not transcription minutes. Pricing as listed: per-second GPU compute (see Modal pricing). Directory tags: commercial-api, model-host. Last vendor-page check: 2026-05-12.
Watch out for: Not a turnkey ASR API — you bring your own model code and pay for compute, not transcription minutes.
Install / use
pip install modal && modal deploy whisper_app.pyFeatures
| Speaker diarization | No |
| Word-level timestamps | No |
| Streaming / real-time | No |
| Languages supported | None |
| HIPAA eligible | No |
Modal (ASR endpoints) vs Whipscribe
| Feature | Modal (ASR endpoints) | Whipscribe |
|---|---|---|
| Category | Transcription APIs | Transcription APIs |
| Pricing | per-second GPU compute (see Modal pricing) | free beta |
| Speaker diarization | — | Yes |
| Word timestamps | — | Yes |
| Streaming | — | No |
| Languages | — | 99 |
| Platforms | API | Web, API, MCP |
Alternatives to Modal (ASR endpoints)
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.