# Qwen3-ASR Docker

> Use this tool when you need to deploy a scalable speech recognition system supporting multiple languages and dialects. It solves problems related to audio transcription, language support, and GPU resource management, providing a user-friendly web UI and REST API for easy integration. The Qwen3-ASR Docker takes in audio files or streams and outputs transcribed text with word-level timestamps, ideal for applications requiring real-time speech recognition and transcription.

Canonical page: https://skillsregistry.net/skills/neosun100-qwen3-asr  
JSON: https://api.skillsregistry.net/v1/skills/neosun100-qwen3-asr

## Description

Provides an all-in-one Docker deployment for Alibaba's Qwen3-ASR speech recognition models supporting 52 languages and dialects. Includes a fastmcp-based server with tools for audio transcription, status checking, language listing, and GPU memory management. Features a dark-theme web UI, REST API with Swagger docs, word-level timestamps via forced alignment, and streaming transcription via WebSocket.

## Trust

- **Trust score (0–1):** 0.63
- **Verification tier:** scanned
- **Last scanned:** 2026-09-02

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** mcp-remote
- **Runtime environment:** api
- **Category:** cloud-infra
- **Updated:** 2026-09-02

## Source

- **Source listing:** [PulseMCP](https://www.pulsemcp.com/servers/neosun100-qwen3-asr)
- **Repository:** <https://github.com/neosun100/qwen3-asr>

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "neosun100-qwen3-asr"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/neosun100-qwen3-asr` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/neosun100-qwen3-asr/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt