Whisper ASR Webservice

Whisper ASR Webservice is a self-hosted Docker web service that exposes OpenAI Whisper and faster-whisper automatic speech recognition models through a REST…

  • AI
  • Tools & Utilities
Whisper ASR Webservice is a self-hosted Docker web service that exposes OpenAI Whisper and faster-whisper automatic speech recognition models through a REST API for multilingual speech transcription, speech translation, and language identification, letting HTTP-capable applications balance speed and accuracy by choosing a model size.

Platform availability

Available on 1 of 10 platforms

  • Unraid
  • TrueNAS (not listed)
  • Umbrel (not listed)
  • ZimaOS (not listed)
  • Proxmox (not listed)
  • Helm (not listed)
  • Mac (not listed)
  • Windows (not listed)
  • Linux (not listed)
  • Any server (not listed)

Health score?A 0-100 score computed nightly from four factors: maintained (40%: recent commits and steady activity), popular (25%: GitHub stars, log scale), easy to install (20%: platforms, install notes, setup guides, Docker image) and light to run (15%: minimum RAM, ARM support). When the source code has been checked, a fifth factor, security check (20%: automated scan for committed secrets, unsafe desktop-app settings, known-vulnerable dependencies and project hygiene; any serious finding caps it at 20), is added. Factors without data are left out and the rest are rescaled.

50/100 Fair

Maintained
65last commit 103 days ago
Popular
713,371 GitHub stars
Easy to install
351 install platform, install notes, Docker image
Light to run
No data
Security check
103 serious findings, 12 minor, CI

Based on 4 of 5 factors.

Availability: platform listings confirmed today, links checked 53 days ago

GitHub stars
3,371
Open issues
125
Last commit
2026-08-09
Latest release
v1.10.0
License
MIT
Activity
stale

Checked 20 days ago - source: GitHub (ahmetoner/whisper-asr-webservice)

Resources & Compatibility

ARM (e.g. Raspberry Pi) is supported.

Checked 54 days ago - source: Docker Hub image tags and published docs

First-install notes

  • Install ffmpeg first Install ffmpeg, a tool that reads and converts audio files, because Whisper needs it to process your recordings. Without it, transcription can fail or produce no output. source
  • Install the Whisper package Run pip install -U openai-whisper to install the speech-to-text engine the web service uses. Without this package, the app cannot turn audio into text. source
  • Fix a common install error If installing Whisper fails with “No module named 'setuptools_rust'”, install setuptools-rust with pip and try again. This missing helper is a common reason the install breaks. source
  • Choose the model size Set the ASR_MODEL variable to the model you want (tiny, base, small, medium, large, large-v1, large-v2). The model downloads on each container start, so the first launch with a large model takes a while. For English-only audio, the .en variants such as tiny.en and base.en give the best results. source

Fetched 37 days ago - each note links to its source

Alternatives

No alternatives collected yet.

Common questions

Does the container download the ASR model on every startup?

Yes, the ASR model is downloaded each time the container starts, and using the large model can make startup slow. You can reduce startup time by storing /root/.cache/whisper on persistent storage.

Have users reported GPU problems with the container?

Yes, an Unraid user reported that the Docker container was not using an Nvidia GPU and asked for troubleshooting help.

Can Whisper hallucinate text during long silences?

Users have reported that long silences can cause Whisper to hallucinate text. A VAD may help, but the thread does not show a clear setup path for Bazarr and whisper-asr-webservice.

Does the service support only OpenAI Whisper?

No, it supports multiple ASR engines, including OpenAI Whisper, Faster Whisper, and WhisperX.

Answers sourced from forums.unraid.net, github.com, selfhost.directory, www.reddit.com

Community

Collected 53 days ago - public community threads