Stop describing bugs.
Point at them.
Your PM clicks the broken thing on the real site and types one sentence. Your developer gets the screenshot, the exact element, the URL and the browser. No screen recordings, no “works on my machine”.
One script tag · Reviewers never pay · Flat rate, not per seat
Founders, PMs, QA, clients
- Click any element on the live site and describe what’s wrong
- Record a session — console and network captured with the replay
- Watch status move from open to resolved without asking anyone
- You are the only one who can mark it approved
Developers
- Open the dashboard and read a complete repro, not a vibe
- Copy the whole bundle as markdown into your editor
- Or let your AI agent pull it over MCP and fix it directly
- Push it to GitHub as an issue when it needs a paper trail
Three steps, and one of them is optional
Drop the script tag
One line in your app. A round Jot button appears on the page edge — under 3KB until someone actually opens it.
They point and type
The cursor becomes a crosshair. They click the broken element, write a sentence, and hit send. Jot captures everything else silently.
You fix it, or your agent does
The Jot lands in the dashboard in realtime. Fix it by hand, or point Claude Code at the MCP server and let it work the queue.
Your agent reads the bug,
fixes it, and reports back.
Every Jot is a complete repro bundle over MCP: body, selector, URL, browser, screenshot, replay. Agents can update status and reply — but approval belongs to the human who reported it. Always.
Flat rate. Reviewers never pay.
Priced by workspace, not by seat — invite the whole client.
Team
- 5 projects · unlimited Jots
- GitHub, Linear, Jira, Trello, Teams
- MCP for AI agents · No branding
Agency
- Unlimited projects & Jots
- AI triage & clustering · Analytics
- Client-ready workspaces
<script src="https://cdn.jsdelivr.net/npm/@usejot/widget@1/dist/jot.js" data-jot-key="jot_pk_your_key"></script>