Ratatosk docs
This page covers every way to consume the data Ratatosk produces.
Ratatosk reads the release notes of cloud native projects every hour and keeps only what needs action. There are six ways to use the results. All six are free, and the web, RSS, API, and MCP need no account.
| Method | Best for | Start here |
|---|---|---|
| Web | Browsing, search, release details | /releases |
| Email alerts | A weekly summary, only when your stack needs action | /settings |
| RSS | Following the full stream in your reader | Alerts and feeds |
| Personal RSS | Just your stack, in your reader | Alerts and feeds |
| REST API | Querying changes (Understanding the data) from scripts and automation | REST API |
| MCP server | AI agents using the data as tools — connect the hosted URL or install your own | MCP server |
Getting started
- For reading: browse recent releases at /releases, then register the projects you want to follow as your stack at /settings. Your stack drives email alerts and your personal RSS feed — see Alerts and feeds.
- For your agent: to start without installing anything, register the
hosted MCP endpoint
https://ratatosk.io/mcpwith your agent. It takes five minutes — see MCP server. - For scripts and automation: a single
curl https://ratatosk.io/v1returns the full endpoint index — see REST API. Agents can also request markdown instead of HTML (/md) and discover the API automatically (/.well-known) — see "Other machine-readable endpoints" in the REST API doc.
Documentation
- About Ratatosk: what the service does and why
- Understanding the data: changes, the three axes, severity, and what "action needed" means
- REST API: /v1 endpoints, examples, limits
- MCP server: hosted and self-hosted setup, and the privacy design
- Alerts and feeds: weekly email alerts, public RSS, personal stack RSS