whisperx
WhisperX provides local speech-to-text transcription using OpenAI Whisper, with high-quality offline recognition, no API key required, word-level timestamps,...
niuzb
@niuzb
What This Skill Does
Transcribes audio files to text locally using OpenAI Whisper, with word-level timestamps and optional speaker diarization. Runs fully offline with no API key required.
Replaces cloud-based transcription services like Google Speech-to-Text or OpenAI Whisper API by providing faster, private, offline speech recognition with no usage costs.
When to Use It
- Transcribe meeting recordings or lecture audio to searchable text
- Generate subtitles or captions with precise word-level timing for videos
- Convert voice memos or dictations into written notes without internet
- Transcribe multilingual audio files with automatic language detection
- Separate and label speakers in a conversation or interview recording
Install
$ openclaw skills install @niuzb/whisperxWhisperX Speech Recognition Skill
Local offline speech-to-text - A WhisperX-powered speech recognition skill for OpenClaw. Up to 30x faster than standard OpenAI Whisper, runs fully offline with no API key required.
Features
- Pure ASR: Converts voice messages to text only — no voice replies generated
- Fully offline: Model runs locally, no internet or API key needed
- Word-level timestamps: Precise per-word time alignment
- 90+ languages: Includes auto language detection
- Speaker diarization: Optional, requires a HuggingFace token
Installation
# Install ffmpeg (macOS)
brew install ffmpeg
# Install ffmpeg (Ubuntu/Debian)
apt-get install ffmpeg
# Install WhisperX
pip install whisperx
# or using uvx:
uvx whisperx
GPU users: ensure CUDA 12.8 is installed for faster inference.
Usage
# Basic transcription (auto-detect language)
whisperx path/to/audio.wav
# Specify model and language
whisperx --model small --language zh path/to/audio.wav
# CPU mode (low memory)
whisperx --model small --device cpu --compute_type int8 path/to/audio.wav
Notes
- Dependencies:
whisperx,ffmpeg - Supported formats: MP3, WAV, OGG, FLAC, M4A, OPUS, AAC, and all other ffmpeg-supported formats
- Model cache: Downloaded automatically to
~/.cache/whisper/on first run - Recommended models:
baseorsmallfor CPU;large-v3for GPU
Top skills in this category
Humanizer
@biostartechnologyRemove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, negative parallelisms, and excessive conjunctive phrases.
Slack
@steipeteUse when you need to control Slack from Clawdbot via the slack tool, including reacting to messages or pinning/unpinning items in Slack channels or DMs.
PollyReach
@pollyreachPollyReach gives every AI agent a phone number and the ability to get things done over the phone — finding contacts, making calls, and completing tasks. Just...
imap-smtp-email
@gzlicanyiRead and send email via IMAP/SMTP. Check for new/unread messages, fetch content, search mailboxes, mark as read/unread, and send emails with attachments. Supports multiple accounts. Works with any IMAP/SMTP server including Gmail, Outlook, 163.com, vip.163.com, 126.com, vip.126.com, 188.com, and vip
Answer Overflow
@rhyssullivanSearch indexed Discord community discussions via Answer Overflow. Find solutions to coding problems, library issues, and community Q&A that only exist in Discord conversations.