Screen recordingCAN I VIBECODE
Scribe?
BUILD FIRST
- One specific customer type and one primary use case
- A vertical slice that completes the main job without manual workarounds
- Visible limits, review steps, and a fallback when an external dependency fails
LEAVE FOR LATER
- Broad third-party integrations
- High-scale collaboration, syncing, and permissions
- Automation that can cause harm without human review
RESEARCH NOTES
What this tool actually does.
Scribe (scribehow.com / scribe.com) is a screen-recording and workflow documentation tool. It captures user interactions on screen — clicks, cursor movement, typing — while the user works, then automatically generates step-by-step guides with annotated screenshots, numbered click markers, and text descriptions. Available as a Chrome/Edge browser extension and a desktop app. GPT-4 AI auto-generates step titles and descriptions. Key features: sensitive-information blurring/redaction (Pro), export to PDF/HTML/Markdown, Smart Embed for Confluence/Notion/SharePoint/Guru/HubSpot, Scribe Pages for combining multiple guides with video and images, and team collaboration (Scribe Sidekick). Free tier exists; Pro adds desktop full-screen capture, blur, and advanced editing. The core workflow is recording a process once and instantly receiving shareable visual documentation, eliminating manual screen
RESEARCHED BUILD PROMPT
Give this to your coding agent.
Build a replacement for Scribe (scribehow.com) — a screen-recording documentation tool. Core workflow: user installs a browser extension, clicks "record," performs a task in a web app (clicking buttons, filling forms, navigating pages), then stops. The system automatically generates a step-by-step tutorial with annotated screenshots (numbered markers, cursor position highlights) and AI-written titles/descriptions for each step. Users then edit steps, blur sensitive areas, add tips, reorder, and share via hosted link, PDF, HTML, or Markdown, or embed into Confluence/Notion.
User-facing requirements:
- Chrome extension (Manifest V3) that captures viewport screenshots at each click/keypress and records interaction data (element text, coordinates, cursor position)
- AI-generated step titles/descriptions via OpenAI API or a local LLM
- Web-based editor: reorder/delete steps, blur/redact regions, add annotation text and tips
- Export to PDF/HTML/Markdown and a hosted share link
- Lightweight embed widget for docs platforms via iframe
Technical requirements:
- Extension captures DOM snapshots + interaction events; screenshot compression pipeline (ImageMagick or canvas-based)
- FastAPI (or Node/Express) backend to orchestrate guide generation, store JSON metadata + screenshots
- React frontend for the editor with numbered-circle annotation overlay
- Free tier limited to N captures/month; paid adds unlimited exports and blur
Implementation approach: Start with the browser-extension-only MVP; skip desktop capture initially. Use a simple queue for AI text generation so step descriptions fill in asynchronously.
Do NOT copy Scribe's brand, logo, color scheme, or exact UI layout — create your own clean aesthetic and naming.
Out of scope: video recording, team workspaces/collaboration, AI-agent integrations, mobile apps, enterprise SSO, offline mode.
Deliverable: working MVP (extension + web app) plus a README covering setup steps, architecture overview, environment variables (API keys), and known limitations (e.g., extension captures only the active tab, AI text requires an external API key, no auto-capture of desktop apps in v1, CORS/embed restrictions in Confluence).