See Vidmatic in action
Circle it. Say it. See it shipped. Watch one change go from a marked-up recording to a merged pull request, with proof at every step.
From Commit to Shareable Demo
One command after you ship. Vidmatic does the other four steps.
You ship
Run /mcp__vidmatic__record in Claude Code. It reads your recent commits to work out what is actually worth showing.
It records
Playwright drives your running app through the real journey in your actual UI. No mock screens or slides.
It narrates
A script is written for what was captured and rendered in an AI voice, timed to the recording that actually happened.
You get a link
The video is encoded and published. Paste the public URL into Slack, a PR, or a release note.
Ship the feature. It demos itself.
Not the demo nobody had time to record. The one that recorded itself.
Nothing To Install
One claude mcp add line, once per machine. No pip install, no local ffmpeg, no test harness to commit. Recording, narration and encoding all run on Vidmatic.
Demos That Match The Build
Each demo is captured from your app as it is the day you ship it, so what people watch is what your users will actually see.
A Link For People Who Never Open Your Repo
Support, sales and your PM get the two-minute version, narration and transcript attached. No screen-share to schedule.
Two Commands, One of Them Once
One connects Claude Code to Vidmatic. The other is the one you run every time you ship.
- 1Connect Claude Code (once per machine)
claude mcp add --transport http --scope user vidmatic https://api.vidmatic.ai/mcpNo key in it: run /mcp, choose Authenticate, and sign in from your browser.
- 2Then, whenever you ship something
/mcp__vidmatic__recordRecords your real UI, narrates it, returns a public link.
The same connection also serves /mcp__vidmatic__audit_ux for a narrated UX critique of any live page, with each finding circled on screen. Nothing extra to install.
One Loop, Three Jobs
Video-to-agent is the whole loop: you record, your agent reads it, files one issue, ships it, and proves it on video.
- Screen recording for AI coding agentsYour agent reads the transcript, every mark you drew and the frames you circled, with your code open.
- Screen recording to GitHub issueOne grounded issue with numbered requirements, each tied to the second you said it.
- AI agent proof-of-work demosA narrated video of the shipped change, recorded from your running app, linked in the PR.
Connecting another tool? Start at the screen recording MCP server page.
Built For Your Workflow
The same recording, read by your agent, for every team that files or reviews work.