uaysk

uaysk / qwen3-asr-openai

Public

qwen3 asr server for openai compatible API

16
2
100% credibility
Found Mar 13, 2026 at 16 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A local server that mimics OpenAI's audio transcription service using Qwen3 speech recognition models, offering a web demo for microphone input, file playback, and full file processing with timestamps.

How It Works

1
🔍 Discover the audio transcriber

You find this free tool that turns spoken words in audio files or live microphone into written text, just like magic.

2
💻 Prepare your computer

Run a simple setup script that checks your hardware and gets everything ready, downloading what it needs automatically.

3
🚀 Start the service

Launch it with one easy command, and it begins listening on your computer for audio to transcribe.

4
🌐 Open the web demo

Go to the web page it opens in your browser, where you pick languages and see simple buttons to start.

5
Choose your way to transcribe
🎙️
Live microphone

Click to speak into your mic and watch words appear instantly as you talk.

▶️
Play file live

Select an audio file to play back and get streaming text as it goes.

📁
Full file transcribe

Upload a recording to get the complete text with time stamps for each part.

✅ Enjoy your transcripts

See the full, accurate text from your audio, with options for details like word timings, ready to copy or save.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 16 to 16 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is qwen3-asr-openai?

This Python repo runs a local Qwen3 ASR server mimicking OpenAI's Whisper API, using the Qwen3-ASR-1.7B model from HuggingFace for accurate speech-to-text. Upload audio files to POST /v1/audio/transcriptions for offline results with optional timestamps, stream progressive output, or pipe live mic data via WS /v1/realtime—plus a demo UI at root for instant testing. It handles multilingual audio like Korean-English mixes locally, dodging cloud costs and latency.

Why is it gaining traction?

Open source qwen3 asr flash model support means no vendor lock-in or qwen3 asr pricing surprises, with smart GPU detection in install.sh for seamless qwen3 asr download and offline runs. Realtime WebSocket and timestamped verbose_json outputs beat basic wrappers, letting OpenAI clients swap in qwen3 asr local without code changes. Devs grab it from this qwen3 github repo for HuggingFace-backed ASR that just works on consumer hardware.

Who should use this?

Backend devs building voice apps or podcasts needing qwen3 asr huggingface integration without API keys. Frontend teams prototyping mic-based UIs via the bundled demo or realtime endpoint. ML tinkerers evaluating qwen3 asr opensource models locally before production.

Verdict

Solid starter for local qwen3 asr api at 13 stars and 1.0% credibility—docs shine with curl examples and run.sh flags, but watch for edge cases in early maturity. Try it if you want free qwen3 asr model inference now; skip for battle-tested scale.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.