Your AI agent can now record your screen, edit it, and publish it

PandaStudio can now hand its screen recorder to an AI agent. The agent starts a real, high-quality recording, captures whatever you point it at, and gets back a video project it can edit and publish, with no recorder window to click and no one watching. Screen recordings you used to sit down and make by hand can now be made for you, on a schedule or on a trigger. This post is where the idea came from, what it does, and the workflows it unlocks.

Where this came from

A founder wrote in with a specific wish, paraphrased here and with the details changed to keep him anonymous:

"I want Claude to record a tutorial of my SaaS for me. It can already drive Chrome, but its own screen capture is low resolution. So I want it to start a proper screen recording, stop it when the walkthrough is done, and then build the final video, adding a voiceover and so on."

His agent already knew how to click through his product. The gap was everything around it: capturing that flow at real quality, and turning raw footage into a finished, narrated video without a person doing the editing. That gap is what we closed, and it turns out to unlock a lot more than one tutorial.

What shipped

PandaStudio already had a high-quality native screen recorder, the same one behind the app's record button. It captures your display through the operating system's own pipeline (ScreenCaptureKit on macOS, Windows Graphics Capture on Windows), not a browser tab. The new part is letting an agent drive it. Three commands:

recording.start begins a recording, the whole primary display by default, or a specific display or window you point it at.
recording.stop ends it, finalizes the video file, and hands back a ready-to-edit project.
recording.list-sources lists the displays and windows available to target.

Pair that with the rest of the agent surface, PandaStudio also connects directly to your own ElevenLabs account for narration (cloned voices included), and can publish straight to YouTube, and the agent owns the whole chain: record, narrate, edit, publish.

What you can automate with it

Once an agent can record the screen on its own, a whole category of videos that used to need a human stops needing one. A few of the most useful:

A release demo video, every time you ship

Wire it into your ship process. When a feature merges, your agent opens the app, records the new flow end to end, narrates what changed in a consistent voice, and publishes the clip to YouTube (unlisted by default) or drops it into your release notes. A "what's new in v2.4" video that makes itself, so shipping a feature and announcing it become one step.

Product tutorials and onboarding

The request that started this. Point the agent at a feature, let it click through, and it produces a narrated how-to. Because it is scripted, you can refresh every tutorial the day your UI changes, instead of re-recording a library by hand.

Changelog and social clips

Turn each release into a short vertical clip for LinkedIn or X, captions and zooms included, without booking screen time. The same recording can become a long-form YouTube walkthrough and a 30-second social cut from one prompt.

Bug reproductions and support answers

Have the agent reproduce a reported bug on screen and attach the recording to the ticket, or record a quick "here is how you do X" to answer a support question with a video instead of a paragraph.

The full loop, end to end

Here is the exact sequence, the way the agent runs it. Every line is a real command against the local PandaStudio API.

1. Record the walkthrough

Start, drive the app, then stop. Stop returns a project you can edit right away.

# optional: see what you can record
pandastudio recording.list-sources --json

# start (whole primary display) — or pass --source="window:123:0"
pandastudio recording.start --json

#  ...the agent now clicks through the feature...

# stop: finalizes the MP4 and creates an editable project
PROJECT=$(pandastudio recording.stop --name="Feature walkthrough" --json \
  | jq -r '.data.projectId')

2. Clean it up

pandastudio transcript.transcribe --id=$PROJECT
pandastudio transcript.remove-silences --id=$PROJECT --minSilenceMs=500

3. Narrate it, in your own voice

Generate narration through your ElevenLabs account, using a voice you have cloned, and drop it on the timeline.

VO=$(pandastudio media.generate-narration \
  --model=elevenlabs-direct \
  --voice="My Cloned Voice" \
  --text="Here is how you connect your first data source." \
  --json | jq -r '.data.audioPath')

pandastudio project.add-audio --id=$PROJECT --path="$VO" --startMs=0

4. Point the eye, then publish

Add a zoom on the moment that matters (a click, a setting), export, and publish. Screen recordings keep cursor data, so zooms can even follow the pointer.

pandastudio project.add-zoom --id=$PROJECT --atMs=8000 --durationMs=1800 --depth=3
pandastudio export.start --id=$PROJECT

# optional: straight to YouTube (defaults to unlisted)
pandastudio export.publish-youtube --id=$PROJECT --title="What's new: data sources"

Why this works without a person in the loop

The capture runs in PandaStudio's background process and writes the video straight to disk. No recorder window, no preview to confirm, no save dialog. That is what makes it safe to hand to an agent: recording.start returns as soon as capture is live, and recording.stop returns the finished project. The agent never needs a screen of its own.

It is also local. The recording, the transcript, and the edit all stay on your machine. The only things that leave are what you choose to send: the short script you synthesize with ElevenLabs, and the final video if you publish it.

One-time setup notes

Screen Recording permission. The first recording asks your operating system for Screen Recording access, the same one-time approval the app's record button uses. If you have ever recorded in PandaStudio, it is already granted and the agent path just works. This is an OS approval, so you grant it once yourself; an agent cannot grant it for you.
ElevenLabs key (optional). Add it once under Settings, Integrations, and your cloned voices show up automatically. Prefer to keep everything offline? PandaStudio also has an on-device narration voice that needs no key.
No microphone needed. The agent path records screen only. Record clean, then let the voiceover carry the narration, so you can rewrite the script without re-recording.

Try it yourself

PandaStudio runs locally on your Mac or PC and connects to Claude, Cursor, and other agents over a localhost API. Have the agent record your product once, narrate it in your own voice, and publish, without opening the editor.

Download PandaStudio Connect your AI agent