Turn YouTube, Reels, podcasts, or local files into transcripts, on-screen OCR notes, chapters, and summaries across web, CLI, REST API, and MCP.
Visit ScribizTested hands-on by Awesome Tools ·
Built for creators, researchers, and coding agents that need both spoken transcripts and on-screen code or slide text from any video.
Starts free with a 2-hour guest video, 30 min/mo, 13 MCP tools, and non-expiring packs from $3.99. Main limits are no live meeting bot and a macOS-tested CLI.
Free plan ($0): 1st video up to 2h with no login, then 30 min/mo and 30 YouTube caption credits. Pro is $10/mo ($84/yr) for 600 min/mo (14-day refund). Non-expiring packs from $3.99 (200 captions) or $5 (300 min).
Disclosure: Scribiz is part of our own product portfolio. Every workflow step, pricing tier, and policy detail below comes from live verification.
Scribiz is a video and audio transcription platform built by Ilias Ism under Superleuk LLC. It turns YouTube links, TikToks, Instagram Reels, podcasts, and local files into speaker-labeled transcripts, chapters, summaries, and on-screen OCR notes. Use it in the browser with no login, via an npm CLI, over a REST API, or through a hosted 13-tool MCP server.
Most transcript tools stop when YouTube captions are disabled or a presenter flashes code without reading it aloud. Scribiz pairs speech recognition with frame-level OCR so you capture the whole video in one pass:
https://scribiz.com/mcp to search moments and quote timestamps.npm install -g scribiz to process local MP4, MP3, or M4A files with Scribiz minutes, BYOK Gemini/OpenRouter, or local models.We walked through the live web workspace, interactive lecture demo, Watch OCR view, bulk YouTube batch flow, MCP and CLI docs, privacy terms, and live pricing tiers.
Paste a link or drop a file and Auto mode picks the cheapest path that works. On YouTube videos with captions, it pulls clean text at 0.1 minutes per video minute (or 1 caption credit). When captions are disabled, it switches to Listen mode (ElevenLabs Scribe v2 with Gemini fallback), quotes the exact cost up front, and labels speakers.

Every result page includes clickable chapter pills, a transcript search bar, and paragraph or subtitle views. One-click Copy for AI bundles summary, chapters, screen notes, and transcript into one prompt block. Exports in SRT, VTT, TXT, Markdown, and JSON are free with zero paywalls.
Standard subtitle scrapers miss anything shown only on screen. Watch mode (1 minute per video minute) scans visual changes and logs slide headings, formulas, tables, and code on an On screen tab. Run Both at 2 minutes per video minute to capture speech and visuals in one pass.

In our live check of the demo lecture and Fireship's Docker in 100 Seconds, Scribiz separated spoken lines from slide callouts and filtered to scenes with text. That makes it practical for lecture notes, silent Loom bug reports, and fast technical tutorials.
For content research and RAG pipelines, the bulk tool accepts video URLs, playlists, and channels up to 1,000 videos per batch. You preview the deduplicated video list and maximum credit cost before starting.

One caption credit covers one YouTube video at any length, and failed videos or repeat downloads cost 0 credits. Finished batches download as a combined Markdown or TXT document, or as a ZIP of individual SRT, VTT, TXT, and JSON files.
Connect AI assistants to https://scribiz.com/mcp with no key for your first 2-hour video, or pass a Bearer key for all 13 tools. Agents call get_video_context first (under ~2,000 tokens), locate moments with search_video, and pull exact windows with get_transcript.

Every MCP response wraps third-party video text inside UNTRUSTED VIDEO TEXT markers to reduce prompt-injection risk. For local runs, npm install -g scribiz supports scribiz login (hosted minutes) or scribiz setup (BYOK Gemini at $0 Scribiz markup).

A REST API at https://scribiz.com/api exposes the same asynchronous job pipeline with OpenAPI 3.1 docs and idempotency keys. Every docs page also serves raw Markdown (.md) alongside llms.txt for coding agents.
Scribiz never uploads a full local video file from the browser; it extracts audio on your device and deletes temporary media after each job. Guest results expire after 24 hours, signed-in results stay private for 30 days, and Scribiz never trains on your content.
Weigh four clear trade-offs before committing:
ffmpeg, and yt-dlp for links).Your first video up to 2 hours is free with no account. A free account adds 30 Listen/Watch minutes per month, 30 one-time YouTube caption credits, and 30-day history.

Pro costs $10/month or $7/month billed yearly ($84/year) for 600 minutes per month, 6-hour videos, and a 14-day full refund on your first payment. Light users can stay on Free and stack non-expiring packs: 300 minutes for $5 ($25 for 2,000; $100 for 10,000). Bulk YouTube caption packs start at $3.99 for 200 videos ($11.99 for 1,200; $23 for 5,000).
Pick Scribiz if you pull transcripts from technical YouTube tutorials, uncaptioned videos, slide-heavy lectures, or playlists. It stands out when you want free exports, visual OCR, non-expiring credit packs, or direct MCP access in Claude and Cursor. Skip it if you need a live bot for Zoom and Google Meet calls, or a native Windows CLI today.
Software engineer and founder (15+ years building web apps). Every listing is verified against live product checks, official pricing pages, and real usage limits before publication.
Updated 16d ago