Picorn: a registered MCP server for editable image and video projects #1534
cueqzapper
started this conversation in
Show and tell
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Picorn is now active in the Official MCP Registry as
io.github.cueqzapper/picorn(v2.2.0).The problem it targets is the flat-output dead end in visual agent workflows. An agent can make an image, but if one crop, headline, or layer is wrong, the usual fallback is another full generation.
Picorn's MCP creates native image documents and video timelines instead:
It is a remote Streamable HTTP server at:
No API key or Picorn account is required for the fair-use MCP path. The endpoint currently exposes 30 tools.
Codex:
Claude Code:
Cursor:
{ "mcpServers": { "picorn": { "url": "https://picorn.com/mcp" } } }A useful first test is:
Current boundaries: it will not reproduce every reference image pixel-perfectly, and I still expect one human art-direction pass. Publishing is kept behind explicit approval rather than being triggered silently by the agent.
Registry entry: https://registry.modelcontextprotocol.io/v0.1/servers/io.github.cueqzapper%2Fpicorn/versions/latest
Tool list, setup, and example workflow: https://picorn.com/mcp?utm_source=github_mcp_registry&utm_medium=showcase&utm_campaign=mcp_showcase_2026_08&utm_content=registered_server_showcase
Disclosure: I build Picorn. I would especially value feedback on the boundary between agent-side editing and the browser-editor handoff: which layer or timeline operations are still missing in real client workflows?
All reactions