Video2Blog
Videos into articles you can verify.
A Cloudflare SaaS that turns a video into a structured article where every paragraph cites the moment in the source it came from.
Video2Blog reads what is said and what is shown in a video: audio is split and transcribed, key frames are extracted and read with OCR, and both become one evidence timeline. The writer only sees that evidence and must cite it, so each paragraph carries timestamp chips that jump back to the source.
The app runs on Cloudflare Workers (Hono API, D1 with full-text search, Workflows, Workers AI) with a Preact front end. It includes accounts, an article editor with autosave and AI rewrites that keep citations, quality checks for uncited claims and unsupported numbers, an SEO pack, exports (Markdown, HTML, WordPress blocks, JSON-LD), a searchable library, public share pages and an admin console. An optional Python engine handles scene detection, OCR and Whisper transcription.
- Type: SaaS application
- Status: In development — Features are covered by unit and end-to-end tests locally. Not deployed yet: the Worker deployment is pending, billing is designed but not wired, and the legal pages are drafts.
- Code: simonvildonelohim/video2blog (private repository)
- Year: 2026
Built
- Paste a link or upload a video
- In-browser media preparation: audio parts, key frames, OCR
- Evidence timeline of speech and on-screen text
- Article with timestamp citation chips
- Quality checks: uncited claims, non-verbatim quotes, unsupported numbers
- Editor with autosave, versions and AI section rewrites
- Exports: Markdown, HTML, WordPress, JSON-LD, zip
- Library with full-text search
- Admin console with usage and jobs
Technology
Cloudflare Workers, Cloudflare D1, Cloudflare Workflows, Workers AI, Hono, Preact, Vite, TypeScript, Tesseract.js, WebCodecs, Python, faster-whisper, Playwright, Vitest
