Captions & VOWords that ride
with the demo.
Generate captions, burn them in or export .srt, then optionally speak them with AI voiceover, so mute autoplay still teaches.
Works with vibe coding agents over MCP
Cursor
Claude Code
Windsurf
Trae
Antigravity
CodexThe problem
Every demo starts from scratch.
Re-doing captions & vo by hand on every take
Opening a full editor for a thirty-second demo
Agents that capture clicks but leave the polish to you
The fix
Captions & VO is built into the studio, not bolted on later.
01Session captions
Refresh from the take, edit cues on the timeline, toggle burn-in vs overlay, font scales with export.
02Download .srt
Hand timing to YouTube or Notion without re-transcribing. Polish stays tied to your cut.
03Translate & speak
Pick speaking and voice languages, then generate AI voiceover clips that follow the cue list.
In the studio
The real Screen Beaver UI, Captions & VO selected.
Rail, inspector, stage, and timeline, the same chrome you get in the Windows desktop app.
Set up once & you’re done.
Install the Windows app first, then MCP, then your agent.
01 Install the app
Download Screen Beaver for Windows. MCP only works after the software is installed and running.
Microsoft Store → install → leave open
02 Wire MCP
Install the CLI, then point any MCP-capable coding agent at Screen Beaver.
$ npm i -g @screenbeaver/mcp
03 Ship the demo
Captions & VO lands on the timeline, export locally.
Same demo. Different finish.
Toggle captions & vo on, watch Screen Beaver do the polish work between a raw take and a shippable cut.
Studio · Captions & VO
On
Portable. No lock-in. Local export.
On the timeline
Nudge, delete, or regenerate, nothing locked after the take.
Delete anytime
Remove markers you don’t want. Your project stays yours.
Export locallyNEW
H.264 MP4 on your machine. No cloud encode required.
One studio · every demo
Stop scrubbing for captions & vo.
Free to start. Wire MCP once, then ask your agent to film while captions & vo lands on the timeline.