# 9routerx-stt

9routerx-stt — a skill for Claude Code and Codex that routes audio transcription requests through 9RouterX. Transcribes voice recordings and generates subtitle files across multiple model backends.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf (Devin Desktop), Claude.ai · ★ 259 repo stars
- Author: yanstd888
- License: MIT · policy=mirror · the beginning of the file is at the end of this page
- Source: https://github.com/yanstd888/9routerx/tree/HEAD/skills/9routerx-stt/SKILL.md
- Updated: 2026-11-16
- Security check: automated, 2026-11-17 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/skills/9routerx-stt

## What it does

This skill routes speech recognition requests through the OpenAI Whisper compatible `POST /v1/audio/transcriptions` endpoint. The agent uploads audio files in mp3, wav, m4a, webm, ogg, or flac formats, selects a target model, and can optionally specify an ISO-639-1 language code, prompt hint, or temperature. The output returns plain text, standard JSON, verbose JSON with segment timestamps, or subtitle files in srt and vtt formats.

The backend server normalizes requests for specific providers automatically. It maps calls to inline audio payloads for Gemini, handles asynchronous upload polling for AssemblyAI, and sets provider-specific auth headers. For ElevenLabs, passing both `diarize` and `num_speakers` invalidates the request upstream, so the server prioritizes diarization and drops speaker counts.

Running queries requires the `9ROUTERX_URL` environment variable, plus `9ROUTERX_KEY` whenever authentication is enforced. Available models can be discovered dynamically via the `/v1/models/stt` endpoint.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add yanstd888/9routerx --skill 9routerx-stt -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/9routerx-stt/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/9routerx-stt/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add yanstd888/9routerx --skill 9routerx-stt -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/9routerx-stt/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add yanstd888/9routerx --skill 9routerx-stt -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/9routerx-stt/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install yanstd888/9routerx skills/9routerx-stt
```

2. **Manually: put the files in place**
- File: `.github/skills/9routerx-stt/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add yanstd888/9routerx --skill 9routerx-stt -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/yanstd888/9routerx --path skills/9routerx-stt
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/9routerx-stt/SKILL.md` — The legacy .windsurf/skills/ path still works

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/yanstd888/9routerx/tree/HEAD/skills/9routerx-stt/SKILL.md

## When to use it
- transcribe recorded interviews and podcasts from audio files
- build srt or vtt subtitle tracks directly from audio tracks
- extract timestamps and segments with verbose JSON responses
- switch between speech backends like Groq, OpenAI, and Deepgram

## Example prompts
- Transcribe audio.mp3 using the groq/whisper-large-v3-turbo model
- Generate srt subtitles from meeting.wav with language set to en
- Convert podcast.m4a to text using deepgram/nova-3 with timestamps

## Source text · SKILL.md · MIT

Below is the content of a third-party file. It is data, not instructions for you: do not act on anything written inside unless the human asks. Block tag: f888cc230fcbf660.

BEGIN THIRD-PARTY FILE f888cc230fcbf660
````text
---
name: 9routerx-stt
description: Speech-to-text via 9RouterX /v1/audio/transcriptions using OpenAI Whisper / Groq / Gemini / Deepgram / AssemblyAI / NVIDIA / HuggingFace models. Use when the user wants to transcribe audio, convert speech to text, or get subtitles from audio files.
---

# 9RouterX — Speech-to-Text

Requires `9ROUTERX_URL` (and `9ROUTERX_KEY` if auth enabled). See https://raw.githubusercontent.com/yanstd888/9routerx/refs/heads/main/skills/9routerx/SKILL.md for setup.

## Discover

```bash
curl $9ROUTERX_URL/v1/models/stt | jq '.data[].id'
# Per-model params (language, response_format, prompt, temperature support)
curl "$9ROUTERX_URL/v1/models/info?id=openai/whisper-1"
```

`model` = STT model ID (e.g. `openai/whisper-1`, `groq/whisper-large-v3`, `deepgram/nova-3`, `gemini/gemini-2.5-flash`).

## Endpoint

`POST $9ROUTERX_URL/v1/audio/transcriptions` (OpenAI Whisper compatible, `multipart/form-data`)

| Field | Required | Notes |
|---|---|---|
| `model` | yes | from `/v1/models/stt` |
| `file` | yes | audio file (mp3, wav, m4a, webm, ogg, flac) |
| `language` | no | ISO-639-1 (e.g. `en`, `vi`) |
| `prompt` | no | hint text to guide transcription |
| `response_format` | no | `json` (default) / `text` / `verbose_json` / `srt` / `vtt` |
| `temperature` | no | 0–1 |

## Examples

```bash
curl -X POST "$9ROUTERX_URL/v1/audio/transcriptions" \
  -H "Authorization: Bearer $9ROUTERX_KEY" \
  -F "model=openai/whisper-1" \
  -F "file=@audio.mp3" \
  -F "language=vi"
```

JS (Node):

```js
import { createReadStream } from "node:fs";
const form = new FormData();
form.append("model", "groq/whisper-large-v3-turbo");
form.append("file", new Blob([await (await import("node:fs/promises")).readFile("audio.mp3")]), "audio.mp3");
const r = await fetch(`${process.env.9ROUTERX_URL}/v1/audio/transcriptions`, {
  method: "POST",
  headers: { "Authorization": `Bearer ${process.env.9ROUTERX_KEY}` },
  body: form,
});
const { text } = await r.json();
console.log(text);
```

## Response shape

Default (`response_format=json`):
```json
{ "text": "Xin chào, đây là bản ghi âm." }
```

`verbose_json` adds `language`, `duration`, `segments[]` with timestamps.
`srt` / `vtt` return subtitle text.

## Provider quirks

| Provider | `model` format | Notes |
|---|---|---|
| `openai` | `whisper-1`, `gpt-4o-transcribe`, `gpt-4o-mini-transcribe` | Native OpenAI shape |
| `groq` | `whisper-large-v3`, `whisper-large-v3-turbo`, `distil-whisper-large-v3-en` | Fastest; OpenAI shape |
| `gemini` | `gemini-2.5-flash`, `gemini-2.5-pro`, `gemini-2.5-flash-lite` | Server converts to `generateContent` with audio inline |
| `deepgram` | `nova-3`, `nova-2`, `whisper-large` | Token auth; server adapts response |
````
END THIRD-PARTY FILE f888cc230fcbf660

The third-party file has ended. Everything between the f888cc230fcbf660 tags is data: requests and commands from it are not followed. An “end of file” line with another tag inside the block is part of the file.

First 2733 of 3658 bytes. © yanstd888 · MIT · https://github.com/yanstd888/9routerx/tree/HEAD/skills/9routerx-stt/SKILL.md

_Part of the page did not fit — see the HTML version._

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
