Fun-ASR-Nano-API

OpenAI-compatible Speech-to-Text API powered by Alibaba FunAudioLLM's Fun-ASR-Nano (800M parameters), supporting 31 languages, Chinese dialects, hotword…

  • AI
  • Productivity
  • Tools & Utilities
OpenAI-compatible Speech-to-Text API powered by Alibaba FunAudioLLM's Fun-ASR-Nano (800M parameters), supporting 31 languages, Chinese dialects, hotword boosting, VAD, and streaming transcription. For self-hosted audio-to-text pipelines with switchable model support via MODEL_ID. Requires an NVIDIA GPU with driver 550+.

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.

38/100 Low

Maintained
59last commit 163 days ago
Popular
61 GitHub stars
Easy to install
351 install platform, install notes, Docker image
Light to run
No data

Based on 3 of 4 factors.

Availability: platform listings confirmed today, links checked 22 days ago

GitHub stars
1
Open issues
1
Last commit
2026-06-10
Activity
stale

Checked 7 days ago - source: GitHub (hsiang-han/Fun-ASR-Nano-API)

Resources & Compatibility

Requirements not published yet.

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

First-install notes

  • Use your account username and password to log in Before starting, provide the account username and password in the login request. If they're missing or wrong, the service won't issue the access key needed for later steps, and wrong details return an error code like -9. source
  • Log in once and reuse the access key Call login once when starting and save the token (the access key the service gives you after login) for later requests. Without it, the service will reject calls for your balance, phone number, or SMS code. source
  • Release unused numbers and avoid extra charges If a phone number doesn't get an SMS, use the release action for that number so it doesn't stay assigned and waste your balance. Also skip the apiReturn status-report call unless your account has special permission, because unauthorized calls can deduct balance. source
  • First start & hardware First start downloads the model from Hugging Face. China users: set HF_ENDPOINT to https://hf-mirror.com. Requires an NVIDIA GPU with driver 550+. source

Fetched 47 days ago - each note links to its source

Alternatives

No alternatives collected yet.

Common questions

Can existing OpenAI transcription clients be pointed at Fun-ASR-Nano-API?

Yes. It exposes an OpenAI-compatible speech-to-text API.

What language and dialect coverage does Fun-ASR-Nano support?

The model guide describes support for Chinese, English, Japanese, seven Chinese dialect groups, and 26 regional accents.

Answers sourced from github.com, www.funasr.com

Community

Collected 49 days ago - public community threads