Shipt drives your app like a real user, exercising the surfaces your PR touched, then updates the docs and cuts the hype video. One command in Claude Code, right on your dev box — no flaky staging, CI, or preview environments.
Your agent writes code faster than you can QA it, update the docs, or cut a launch video. Its self-generated tests pass while real bugs slip into production — everything outside code velocity feels like it's breaking.
Shipt helps you ship with quality. It identifies what changed and drives the app like a chaotic user, updates the docs, and cuts a short launch video from the QA session. One /shipt:release command in Claude Code — no new CI env, no new dashboard.
Runs against your dev box. Reads the diff, figures out which flows matter, then clicks through them in a real browser. No Playwright babysitting. No staging to configure.
Diff-aware. Knows what changed in the UI and rewrites the affected pages. Opens a docs PR in parallel with the code one.
Takes the session replay, captions the key moments, cuts it to 18 seconds. Drops it in the changelog channel. Your launch is already filmed.
Two minutes to install. Works on your next PR.