Curio CLI

The curio command gives you a second way into your projects. Anything you would normally do by clicking around a project—finding figures, reading and changing their content, filing them into Organizer sections, bringing content in and sending it back out—you can also do with a command, from a script, or through an AI agent:

• Query. Search a project with Curio’s query language, spanning text, tags, dates, status, ratings, flags, and metadata.
• Read and update. Inspect and change titles, text, checkboxes, tags, dates, meta values, figure geometry, and dozens of other figure and Organizer item fields.
• Organize. Create Organizer sections, folders, and idea spaces, then move figures and Organizer items where they belong.
• Import and export. Bring files, text, and structured content into a project, spread PDF pages across idea spaces, and export figures and idea spaces back out.
• Inspect. List the Organizer hierarchy, project assets, available tags, available fields, Project Center categories, and the state of the running app.
• Navigate. Send the Curio app to a specific figure or idea space, open deep links, or work with whatever is currently selected.

Results print as readable text for people, or as JSON with meaningful exit codes for scripts and agents.

For AI, the CLI ships bundled skill files that teach command line agents such as Codex and Claude Code how to drive the curio command, plus a built-in MCP server so MCP hosts such as ChatGPT Desktop, Claude Desktop, and Antigravity can work with your Curio data directly.