Transcribe Anything — guide skill for captions & subtitles

For AI agents: find the right skill in one call — curl -s "https://skills.varg.ai/api/search?q=<what the person wants>" returns the matching use case and the best skills with install commands, as markdown (Russian works too). Full index: https://skills.varg.ai/llms.txt. Every page has a markdown twin: add .md to its path. Before generating anything, show the person the options on the use-case page.

Transcribe hours of audio or video with whisper, whisperX or cloud STT: word timestamps, speaker labels, SRT/VTT.

• Type: Guide · Captions & subtitles
• Slug: swyx-transcribe-anything
• Status: live
• Publisher: Shawn Wang (swyx) (community)
• License: MIT
• Source: https://github.com/swyxio/skills/tree/main/transcribe-anything
• Works in: Claude Code, Codex, Cursor
• Requires: ffmpeg, At least one ASR backend (openai-whisper or MLX Whisper by default), HF_TOKEN with accepted pyannote terms for diarization (optional), Provider keys only for cloud backends (OPENAI_API_KEY, GROQ_API_KEY, DEEPGRAM_API_KEY, ASSEMBLYAI_API_KEY, GEMINI_API_KEY)
• Cost: Free with local backends; cloud backends bill your own provider keys
• Adoption: 10 installs, 176 GitHub stars (GitHub + skills.sh, 2026-10-26)

What it does

A backend-routing transcription guide from swyx: it prepares media with ffmpeg (audio extraction, silence detection and trimming, chunking under API size limits), then runs local Whisper, MLX Whisper, faster-whisper, whisper.cpp, insanely-fast-whisper or whisperX, or the OpenAI, Groq, Deepgram, AssemblyAI or Gemini APIs. It covers word-level timestamps (word_timestamps=True, whisperX alignment), pyannote or NeMo diarization with a QA gate, custom vocabulary and SRT/VTT export, although its default output is a Markdown transcript. Instructions only; the repo README says the collection reflects the author's own setup, and setup lines use pip3 install --break-system-packages.

Use when

• Transcribe a video, audio file or podcast
• Word-level timestamps to drive captions
• Speaker labels on a long interview
• Very long files (1–8+ hours) that need chunking

Not for

• Long diarization on a local CPU without explicit opt-in (the skill says not to)
• Styling or burning captions (pair with caption-animation or ffmpeg-captions-subtitles)

Install

• Claude Code: npx skills add https://github.com/swyxio/skills --skill transcribe-anything
• Codex: npx skills add https://github.com/swyxio/skills --skill transcribe-anything
• Cursor: npx skills add https://github.com/swyxio/skills --skill transcribe-anything

Models it drives: whisper, whisperx, faster-whisper, gpt-4o-transcribe

More captions & subtitles skills

• 3D Camera Captions: Captions that live in 3D around a talking head: the camera flies through the words, heroes pass behind the speaker.
• Remotion Captions: Transcribe speech and burn in animated captions with Remotion, or import an existing SRT file.
• ElevenLabs Speech to Text: Transcribe audio and video with word timestamps, speaker labels and keyterms, ready for subtitles.
• ElevenLabs Speech to Text (inference.sh): Transcribe audio with ElevenLabs Scribe, speaker diarization and forced alignment, via inference.sh.
• Speech to Text: Transcribe audio with ElevenLabs Scribe v2 or Whisper V3 Large, with timestamps and speaker labels.
• Pretext Text Metrics: Measure and fit text without the DOM using pretext: auto-fit sizes, know line breaks before render, canvas/WebGL text.