Record it. Agents see it.
Free, open-source screen recording CLI + MCP server for AI agents. Zero cost. 100% local. No cloud. No API keys.
- Screen Recording — Capture full screen, windows, or regions via ffmpeg
- Screenshots — Instant PNG capture with JSON metadata output
- MCP Server — 5 tools exposed via Model Context Protocol (stdio)
- Frame Extraction — Scene-detection key frames for vision model analysis
- Event Logging — JSONL event log synced to video clock (clicks, keystrokes, annotations)
- Cross-Platform — macOS, Windows, Linux
- $0 Cost — No APIs, no cloud, no hosting, no subscriptions
# For Claude Code
claude mcp add recit -- npx -y @recit/mcp
# For Cursor / VS Code — add to mcp.json:
{
"mcpServers": {
"recit": {
"command": "npx",
"args": ["-y", "@recit/mcp"]
}
}
}
# Prerequisites (both free)
brew install ffmpeg # macOS
sudo apt install ffmpeg # Ubuntu/Debian
winget install ffmpeg # Windows# Start recording
recit start [--output path] [--fps 15] [--json]
# Stop recording
recit stop <id> [--json]
# Capture screenshot
recit snap [--output path] [--json]
# Extract frames from video
recit frames <video> [--max 20] [--output dir]
# Check recording status
recit status [id]
# Start MCP server
recit mcp| Tool | Description |
|---|---|
recit_start |
Start screen recording |
recit_stop |
Stop recording, get file path |
recit_snap |
Capture screenshot |
recit_frames |
Extract key frames from video |
recit_status |
Get recording status or active count |
AI Agent (Claude/Codex/Cursor)
│
│ stdio (stdin/stdout)
▼
recit MCP Server (Node.js, local)
│
│ child_process
▼
ffmpeg (free, local)
│
▼
.mp4 / .png saved to user's disk
$0 spent. Nothing uploaded. Everything local.
git clone https://github.com/recit-ai/recit.git
cd recit
npm install
npm test # Run 43 tests
npm run build # Build TypeScript| Layer | Tool | Cost |
|---|---|---|
| Recording | ffmpeg | $0 |
| Agent Integration | MCP SDK (stdio) | $0 |
| Distribution | npm | $0 |
| Code Hosting | GitHub | $0 |
| Runtime | Node.js | $0 |
| Testing | Vitest | $0 |
Apache 2.0