15 chapters
Anchor Cards documentation
Your cards are Markdown that an AI agent writes in your own GitHub repository. Anchor reads them to quiz you, and writes back only your review log and a snapshot of what you keep forgetting.
This documentation is the public specification. If it disagrees with the app, one of the two is wrong — either a bug to fix or a page to rewrite. Where iOS, Android and macOS behave differently, it says so.
Start here
- 01What Anchor Cards isA spaced-repetition app that holds no cards of its own. Your cards are Markdown that an AI agent writes in your GitHub repository; the app only reads them and quizzes you.
- 02Getting startedInstall it, try the bundled samples, then connect the GitHub repository your AI agent wrote the cards into, and start reviewing.
- 03Using it with Claude Code or CodexHave a coding agent write the cards in your GitHub repository and fix them through pull requests. On Mac, Anchor can tell the agent which card is on screen.
Cards
Using it
- 06ReviewingThe order cards come in, what the four grades mean, undo and re-run, and how to read the progress bar.
- 07BrowsingPulling up a specific card outside the queue. Search runs entirely on the device, and nothing here is graded.
- 08Changes (pull requests)Read the `.md` files in an agent's open PR — as cards — before you merge.
How it works
Settings and platforms
- 11SettingsWhere your cards live (GitHub), sync, new cards per day, and current status. There is almost nothing to type.
- 12Anchor ProThe only difference is how many new cards you can start per day. One-time purchase; everything else is in the free version.
- 13Privacy and safetyIt talks only to where your cards live — your GitHub repository. No analytics SDK is included. For devices already connected to Anchor Cloud, how it holds data is set out here too.
- 14Platform differencesiOS, Android and macOS read the same cards and the same review log. Where they differ, they differ on purpose.