Edit programme config, add subjects, or export a handoff bundle for a fresh Claude Code session.
Generates a single downloadable markdown file bundling: current programme.json, all subject data, memory hard-rules, library catalogue, session state, and a primed Claude Code prompt. Drop it into any fresh Claude Code session and say "continue" — the new session picks up exactly where we are.
Any cards or quiz items you've edited or added in the app live in your browser's localStorage as a private overlay. They show in the app right away, but to make them permanent (and shareable across devices / sessions), export them as a patch and hand to whoever maintains the canonical content.
Tip: when you give the exported patch to Claude Code, paste: "Apply my edits patch at <path> — register as kind:manual-md in the ingestion ledger and merge into module.json." That preserves provenance.
Preferred playback settings for the unit audiobook player. Apply on next open.
Pre-rendered audiobooks (when available) play at the speed you pick here. The voice picker only matters when the audiobook MP3 isn't bundled — the player falls back to your browser's voices.
The 6-step welcome tour shown on first visit. Replay it any time.
If you're suggesting permanent improvements that classmates might use, pick a handle (a chosen name — never a real one). It travels with your suggestions so the canonical content can credit you.
Per the PRIVACY rule: handles must be chosen, not derived from your real name.
These tools are for moving state between devices, contributing improvements back to the canonical content, or experimenting with the spaced-repetition algorithm. The defaults are sensible — collapse this whole section once you've set things up.
Paste an endpoint URL that supports HTTP GET + PUT (a Gist, a webdav drop-box, an S3 pre-signed URL, …). The app sends your handoff bundle there, and pulls + merges on launch. We host nothing.
⚠ The endpoint sees your full state. Use a private, authenticated URL. Think of this like the handoff bundle, but over HTTP.
SM-2 (default) is the classic Anki algorithm — robust, well-understood. FSRS-6 is a 2024 algorithm that tends to need 20-30% fewer reviews for the same retention by modelling memory stability and difficulty per card. benchmark
Switching seeds the new algorithm from your existing per-card state — your schedule won't reset. You can switch back anytime.
Bundles everything in your browser — settings, SRS schedule, activity log, streak, study focus, edits, notes, suggestions queue, and unit-reviewed flags — into one file. Use this to:
Merge mode: smart per-field merge — SRS takes max progress, activity is unioned, notes are deduplicated by (text, createdAt), edits prefer the more recent _editedAt. Safe to import from an older or newer device.
Replace mode: wipes localStorage and restores from the bundle. Use only on a fresh machine or to recover from corruption.
Apply merge-receipt: tiny file Claude returns after a merge cycle — auto-clears the suggestion-queue items that have been folded into canonical content.
Edits and notes you've flagged with "🚀 Suggest as permanent improvement" live here until Claude folds them into the canonical content. Hand the export to Claude Code and ask for a merge run; once each item is integrated, click "Mark applied" to clear it from your queue.
Workflow: (1) study, flag improvements with the 🚀 checkbox in any edit modal · (2) export this queue and hand to Claude Code with: "Run a merge of my suggestions queue at <path> into the canonical content" · (3) Claude folds each item, returns a list of applied IDs · (4) paste the IDs here to clear them from your queue.
config/programme.json to your browser's downloads. Drop it in config/programme.json to replace the file. (No server-side write — this app is static HTML + JSON.)