Summary
Screenshots stop being things you paste and start being things you can look up.
Gyazo already holds every capture you took; full-text search is what turns that pile into something an agent can retrieve from, and upload closes the loop when the agent produces an image of its own. Organising is not part of the deal — no albums, no deletion, no editing metadata after the fact.
What it is
A four-tool server over the Gyazo API: full-text search across your uploads, capture retrieval with metadata, and upload.
What you get
- Full-text search across the captures uploaded to the account, paged by page and per
- Any capture fetched by its id or its Gyazo URL, image content and metadata together
- The most recent upload fetched without needing to know its id first
- New images uploaded with a title, description, referer URL and originating app
Requirements
GYAZO_ACCESS_TOKEN — search and the latest-upload lookup both run against the captures belonging to that token.
Setup effort
One command plus a key — npx @notainc/gyazo-mcp-server, then supply credentials