API Reference
Programmatic access to your Raster libraries. Query assets, upload images, and build integrations.
Raster is built API-first. Browsing libraries, uploading images, reading and applying tags, and organizing assets are available programmatically, so you can connect Raster to your website, CMS, build pipeline, or internal tools.
Choose an interface
REST API
A resource-oriented HTTP API for your libraries, assets, and tags.
MCP Server
Connect Claude, ChatGPT, Cursor, and other AI assistants to your libraries over the Model Context Protocol.
SDK
Add Raster to your own app or plugin: sign-in, browse, search, upload, and a drop-in React picker.
For AI agents
An AI agent can manage a user's images end to end in Raster, including for a user who has no account yet.
Install the Raster skill
Add Raster to Claude, Cursor, Copilot, and other agents with one skills-CLI command.
Using Raster from an AI agent
A runbook for LLMs: discover Raster, authenticate, and run the upload, tag, search, and hand-off loop.
Start without an account
One call creates a library and key from an email. Upload right away; the user claims it later.
What you can build
- Serve CDN-cached images in several sizes on a website without a traditional CMS.
- Sync a Raster library into a static site generator at build time.
- Push images into Raster from an existing upload flow or automation.
- Pull asset metadata and tags into internal dashboards and search tools.
Before you start
The REST API authenticates with an organization API key or an OAuth access token. Create and manage keys, and grant per-library access, in your organization settings. See Authentication for the full walkthrough.
The MCP server and the CLI also support OAuth, signing you in as yourself so a connection follows your live library access. See MCP authentication to connect over OAuth.
The REST API is in Alpha. Follow the REST changelog for breaking changes, and reach out at support@raster.app with feedback.