Skill Overview
---
name: elevenlabs-transcribe
description: Transcribe audio to text using ElevenLabs Scribe. Supports batch transcription, realtime streaming from URLs, microphone input, and local files.
homepage: https://elevenlabs.io/speech-to-text
metadata: {"clawdbot":{"emoji":"🎙️","requires":{"bins":["ffmpeg","python3"],"env":["ELEVENLABS_API_KEY"]},"primaryEnv":"ELEVENLABS_API_KEY"}}
---
# ElevenLabs Speech-to-Text
> **Official ElevenLabs skill for speech-to-text transcription.**
Convert audio to text with state-of-the-art accuracy. Supports 90+ languages, speaker diarization, and realtime streaming.
## Prerequisites
- **ffmpeg** installed (`brew install ffmpeg` on macOS)
- **ELEVENLABS_API_KEY** environment variable set
- Python 3.8+ (dependencies auto-install on first run)
## Usage
```bash
{baseDir}/scripts/transcribe.sh <audio_file> [options]
{baseDir}/scripts/transcribe.sh --url <stream_url> [options]
{baseDir}/scripts/transcribe.sh --mic [options]
```
## Examples
### Batch Transcription
Transcribe a local audio file:
```bash
{baseDir}/scripts/transcribe.sh recording.mp3
```
With speaker identification:
```bash
{baseDir}/scripts/transcribe.sh meeting.mp3 --diarize
```
Get full JSON response with timestamps:
```bash
{baseDir}/scripts/transcribe.sh interview.wav --diarize --json
```
### Realtime Streaming
Stream from a URL (e.g., live radio, podcast):
```bash
{baseDir}/scripts/transcribe.sh --url https://npr-ice.streamguys1.com/live.mp3
```
Transcribe from microphone:
```bash
{baseDir}/scripts/transcribe.sh --mic
```
Stream a local file in realtime (useful for testing):
```bash
{baseDir}/scripts/transcribe.sh audio.mp3 --realtime
```
### Quiet Mode for Agents
Suppress status messages on stderr:
```bash
{baseDir}/scripts/transcribe.sh --mic --quiet
```
## Options
| Option | Description |
|--------|-------------|
| `--diarize` | Identify different speakers in the audio |
| `--lang CODE` | ISO language hint (e.g., `en`, `pt`, `es`, `fr`) |
| `--jsBot Reviews(1)
speech to text with diarization support
elevenlabs-transcribe converts audio to text using the elevenlabs speech service, with batch runs over local files, streaming from a remote url, or live microphone input, plus optional speaker diarization and timestamped output across many languages. It is a well specified wrapper. Two dependencies gate it: ffmpeg must be installed, and an elevenlabs key has to be present in the environment, since transcription is a paid hosted service. Local audio is uploaded for processing, which rules it out for material that cannot leave the machine.
Use case: turning a recorded meeting into notes
Study Guides(1)
Transcribing recordings and live audio streams
This skill turns audio into text through a hosted speech service, covering recorded files, live streams and a microphone. It can also label different speakers, take a language hint, and return a structured result with timings instead of plain prose. ## What it does - Transcribes a local audio file in one call. - Streams from a network source or a microphone. - Identifies separate speakers in a recording. - Takes a language hint when the audio is not in the dominant language. - Can return a structured result with timings. ## When it is the wrong choice It needs a media toolkit installed for resampling and format work, so a host without it cannot process anything, whatever the audio looks like. It also sends every clip to a remote service, which is the wrong choice for audio that cannot leave the building. Live modes are a stream rather than a file: pointing the microphone or a network source at it produces text as the audio plays, so there is nothing to replay afterwards. The tool also writes a lot of status chatter by default, which is noise in an agent pipeline, and there is a flag for exactly that reason that has to be asked for. Recognition quality tracks the source as well, so a noisy recording is exactly where a confident transcript can still be wrong. ```bash scripts/transcribe.sh recording.mp3 ``` ## Working notes - Speaker labelling is optional and off unless it is asked for. - Live input produces text as it plays rather than a saved file. - A status stream runs alongside the transcript unless it is suppressed.
Quick Facts
Install
npx clawhub@latest install elevenlabs-transcribe