Let Claude Desktop and Cursor actually watch videos (MCP, fully local) My AI agent learned to watch videos this week. Here is the 3-line version of how to give yours the same ability. claude-real-vid
By Coderz Club · 2026-08-03 · Tags: ai
Let Claude Desktop and Cursor actually watch videos (MCP, fully local)
My AI agent learned to watch videos this week. Here is the 3-line version of how to give yours the same ability. claude-real-video (MIT, 1.9k stars) extracts scene-aware, deduplicated keyframes plus a timestamped transcript from any video URL or local file — all processing on your machine. Since 0.8.0 it ships as an MCP server, so MCP clients can request a video directly. Install pip install 'claude-real-video[mcp]' Claude Code: claude mcp add crv -- crv-mcp Claude Desktop — add to claude_desktop_config.json: { "mcpServers": { "crv": { "command": "crv-mcp" } } } Cursor and any other MCP client: same stdio command, crv-mcp. What you get Two tools show up: watch_video(source, max_frames, language, transcribe) — downloads (URL) or reads (local file), extracts keyframes at scene changes, drops near-duplicates, transcribes with Whisper, and returns the transcript plus the first batch of frames as images straight into the conversation. get_frames(source, start_index, count) — pages through the rest of the frames. Analyses are cached per source under ~/.cache/crv-mcp, so a follow-up question about the same video is instant. Why scene-aware instead of 1 fps Fixed-interval sampling wastes context on near-identical frames and still misses fast cuts. Scene detection keeps the frames that actually differ: a 58-second test clip goes from 58 sampled frames to 26 that matter. Fewer tokens, nothing missed. The benchmark folder in the repo has the full comparison, reproducible. Honest notes Verified end-to-end on Claude Code (a real model run described a test video's frames correctly through the tool). Claude Desktop and Cursor speak the same MCP stdio protocol — config above; open an issue if anything misbehaves. Transcription needs Whisper (pip install 'claude-real-video[whisper]' or [fast]); frames work without it. The mcp SDK shipped 2.0 on July 28 and moved FastMCP — 0.8.0 dual-imports both majors, tested separately on each. If you build your own MCP server this month, test in a clean venv. Repo: https://github.com/HUANGCHIHHUNGLeo/claude-real-video
My AI agent learned to watch videos this week. Here is the 3-line version of how to give yours the same ability. claude-real-video (MIT, 1.9k stars) extracts scene-aware, deduplicated keyframes plus a timestamped transcript from any video URL or local file — all processing on your machine. Since 0.8.0 it ships as an MCP server, so MCP clients can request a video directly. Install pip install 'claude-real-video[mcp]' Claude Code: claude mcp add crv -- crv-mcp Claude Desktop — add to claude_desktop_config.json: { "mcpServers": { "crv": { "command": "crv-mcp" } } } Cursor and any other MCP client: same stdio command, crv-mcp. What you get Two tools show up: watch_video(source, max_frames, language, transcribe) — downloads (URL) or reads (local file), extracts keyframes at scene changes, drops near-duplicates, transcribes with Whisper, and returns the transcript plus the first batch of frames as images straight into the conversation. get_frames(source, start_index, count) — pages through the rest of the frames. Analyses are cached per source under ~/.cache/crv-mcp, so a follow-up question about the same video is instant. Why scene-aware instead of 1 fps Fixed-interval sampling wastes context on near-identical frames and still misses fast cuts. Scene detection keeps the frames that actually differ: a 58-second test clip goes from 58 sampled frames to 26 that matter. Fewer tokens, nothing missed. The benchmark folder in the repo has the full comparison, reproducible. Honest notes Verified end-to-end on Claude Code (a real model run described a test video's frames correctly through the tool). Claude Desktop and Cursor speak the same MCP stdio protocol — config above; open an issue if anything misbehaves. Transcription needs Whisper (pip install 'claude-real-video[whisper]' or [fast]); frames work without it. The mcp SDK shipped 2.0 on July 28 and moved FastMCP — 0.8.0 dual-imports both majors, tested separately on each. If you build your own MCP server this month, test in a clean venv. Repo: https://github.com/HUANGCHIHHUNGLeo/claude-real-video