llms.txt Content
# Scribiz
> Paste a YouTube, TikTok or podcast link, or upload a file. Get the transcript, on-screen text, summary and chapters. Export SRT, VTT, TXT, Markdown or JSON.
Scribiz returns the context of any video: a transcript with speakers and timestamps, the text and scenes on screen, a summary and chapters. Input is a link or a file. Output is SRT, VTT, TXT, Markdown, Context (Markdown with on-screen notes) or JSON. It works when a video has no captions. It runs on the web, as a command-line tool (npm install -g scribiz), as an MCP server and as an HTTP API. A Mac app is available to download.
## When to use Scribiz
Reach for Scribiz when you have a video or audio link, or a local media file, and need to know what is in it:
- The transcript of a YouTube or podcast link, a direct media link or a file, with timestamps and speakers, also when the video has no captions. TikTok, Instagram, X, Vimeo and Facebook links work on a best-effort basis.
- What was shown on screen: slides, code, interfaces and text overlays, with the time each appeared (the Watch mode).
- A summary and chapters of a long video, or answers to questions about it that cite timestamps.
- Subtitles to download: SRT or VTT.
- Notes from a local recording (a meeting, a lecture, a screen recording). The command-line tool sends the audio, and for Watch a low-resolution copy, never the original video.
It is not for editing or generating video, live captions during a call, live streams in progress, or DRM services such as Spotify and Netflix.
How to call it:
1. MCP, for Claude, Cursor, Codex, Grok and other MCP clients: connect to https://scribiz.com/mcp (Streamable HTTP). Without a key it reads captions, videos already processed and, for your first video (up to 2 hours), the link itself (never the picture). With an API key (`Authorization: Bearer sbz_live_...`) or OAuth 2.1 (`/.well-known/oauth-authorization-server`, or `https://scribiz.com/mcp?oauth=1` to force the OAuth sign-in challenge) it uses
OpenAPI Spec (preview)
{"components":{"schemas":{"AskRequest":{"additionalProperties":false,"description":"Exactly one of job_id or context.","properties":{"context":{"description":"The Context of a finished job, exactly as it was returned (at most 1 MB). A link or a file reference is not a Context: it is refused with 400.","type":"object"},"job_id":{"type":"string"},"question":{"maxLength":8000,"minLength":1,"type":"string"}},"required":["question"],"type":"object"},"AskResponse":{"properties":{"answer":{"type":"stri