Recorded demos

How it works

Every workspace box carries devbench video. It drives the deployed app with a real browser, draws a cursor and a highlight ring so you can see what is being pointed at, records the session, and burns a camera move and captions into an MP4. Nobody operates it and there is no editing pass: an agent writes a take, and the video falls out of running it.

It exists because of what a green test run cannot tell you: whether the thing you asked for is the thing that was built. A test says the code does not break; only a video says what it does, and it is what lets somebody give feedback without checking the branch out and clicking through it.

A take is content, and it lives in the repo at devbench/takes/<name>.py, committed next to the feature it shows. It reads like a narration script:

def take(t):
    t.goto("/")
    t.mark_ready()                       # everything before this is trimmed off
    t.say("DevBench records a video of every change it finishes.", 4.0)
    t.beat("[data-testid='kanban-board']", "Work is tracked per workspace.", 3.6)
    t.click("a[href^='/workspaces/']", after=3.0)

The one architectural rule: the live page is never transformed. The obvious way to zoom is transform: scale() on the page while it records, and it fails twice, the browser's click coordinates stop matching where elements are, and the app's own responsive layout shifts while it is being filmed, so the video stops showing the app as users see it. Instead the browser captures a flat frame and the take merely logs where the camera should have been; the camera move is a post-process:

   real page (untouched)
        |
        |  Playwright  ------------>  capture.webm   (flat, no transform)
        |
        +- take logs -------------->  timeline.json  {zooms, captions, trim}
                                            |
                    ffmpeg: trim -> zoompan -> drawtext -> H.264
                                            |
                                       <name>.mp4

Two consequences follow. Captions are drawn after the zoom, so a push-in can never magnify a caption and slice it off at the frame edge. And because the capture and the timeline are separate artefacts, re-rendering is free, which matters, because a take that submits a form submits it again for real, so a caption typo must never mean re-shooting.

The finished MP4 is uploaded to the workspace's shared files and can be posted straight into the chat and attached to the task it belongs to, where it plays inline rather than arriving as a link somebody has to decide to open.