The video editorother AIs hire.
EditorBuddy isn't just a website. It's a server: the same orchestrator, the same 12 agents, the same memory of your style — callable by Claude, by your own agents, or by a script.
Add one server, get the whole team
Point any MCP-compatible client — Claude Desktop, Claude Code, your own agent — atmcp.editorbuddy.comand it gets every tool our orchestrator uses internally: transcribe, cut, caption, translate, dub, find clips, publish.
{
"mcpServers": {
"editorbuddy": {
"url": "https://mcp.editorbuddy.com"
}
}
}Or call it directly
Same orchestrator, plain HTTP. Send a source video and a prompt — the orchestrator plans the job, runs it, and streams progress back.
curl -X POST https://api.editorbuddy.com/v1/jobs \
-H "Authorization: Bearer $EDITORBUDDY_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"source_url": "https://files.example.com/podcast.mp4",
"prompt": "Cut 5 shorts with captions and remove silences"
}'One prompt in. A planned graph out.
The orchestrator reads your prompt and source file, then plans which agents run, in what order, and what runs in parallel. This is the same graph whether the request comes from the web app, the REST API, or an MCP client.
Stream the agents working, not just the result
Every job streams agent-by-agent progress over SSE — the same events that drive the activity panel in the web app. A calling agent can narrate the run instead of going quiet for two minutes.
event: agent.update
data: {"agent":"transcription","status":"working","note":"23 scenes"}
event: agent.update
data: {"agent":"shorts","status":"success","note":"5 clips ranked"}
event: job.complete
data: {"job_id":"job_8f2c","outputs":5}Get an API key andgive your agent a video editor.
Sign up, then grab a key in Settings → API & MCP. Full docs are coming to docs.editorbuddy.com — until then, the MCP server and REST examples above cover the core surface.