Recall Garden
unlistedby Recall Garden
Active recall, spaced repetition, AI study packs, and note verification for Markdown notes.
Recall Garden
Recall Garden (忆园) is an Obsidian study system for active recall, spaced repetition, structured Markdown cards, AI-assisted study packs, and note verification.
It is designed for any subject. The plugin does not assume a specific school, exam, profession, or knowledge domain.
Features
- FSRS-6 spaced-repetition scheduling with configurable retention and review limits.
- Structured eight-section study cards for concepts, comparisons, and applied problems.
- Daily review queues, study calendar, exam countdown, diagnostics, snapshots, and recoverable review history.
- AI-generated study packs containing a 30-second answer, missing sections, cloze questions, a four-option discrimination question, and comparison questions.
- A complete preview of every generated item and the final Markdown before anything is written back.
- Safe managed blocks that preserve handwritten content and existing question-bank answer state.
- AI note verification with saved reports and guarded, per-issue corrections.
- No telemetry, advertisements, or automatic AI requests.
Card format
New cards use this frontmatter marker:
---
type: recall-card
subject: Biology
module: Cell Biology
topic: Photosynthesis
---
Recall Garden also recognizes the legacy values study-card, 学习卡, and 名词解释 for backward compatibility.
A review card needs a standard-answer section and a 30-second-answer section. The built-in card creator generates a complete eight-section template.
Installation
Community directory
After the plugin is approved, install Recall Garden from Obsidian's Community plugins directory.
Manual installation
- Download
main.js,manifest.json, andstyles.cssfrom a GitHub release whose tag matches the manifest version. - Place the files in
<vault>/.obsidian/plugins/recall-garden/. - Reload Obsidian and enable Recall Garden under Community plugins.
For pre-directory testing, the repository can also be installed through BRAT.
AI providers
AI features are disabled by default. Users explicitly choose and configure a provider:
- DeepSeek API: uses a user-provided API key stored through Obsidian SecretStorage.
- ChatGPT Codex device login: an optional compatibility channel using a plugin-specific OAuth session. This channel uses an upstream interface that is not a stable public OpenAI API and may stop working when the service changes.
No AI provider is contacted during startup, vault scanning, scheduling, ordinary review, or diagnostics. A request is sent only after the user clicks an AI action.
Read PRIVACY.md before enabling an AI provider.
Data safety
- Review data uses Obsidian's plugin data API.
- API keys and OAuth tokens use Obsidian SecretStorage and are excluded from snapshots.
- Stable card IDs preserve review history across file renames.
- Unknown newer data schemas enter read-only protection instead of overwriting data.
- AI write-back uses
Vault.process, verifies that the source has not changed since preview, and modifies only explicit managed regions or uniquely matched corrections. - Production releases contain no developer Vault path, local deployment configuration, or user data.
Development
Requirements: Node.js 20 or newer and npm.
npm ci
npm test
npm run check
npm run build
The build creates a minified main.js. Obsidian release assets are:
main.jsmanifest.jsonstyles.css
main.js is intentionally not committed to the repository; it is attached to GitHub releases.
Release policy
- The GitHub release tag must exactly match
manifest.json's version, without avprefix. - Source, lockfile, tests, manifest,
versions.json, and release notes are committed. - Built assets are uploaded to the matching GitHub release.
- Schema changes must preserve stable IDs, scheduling history, archived cards, and saved AI questions.
License
Recall Garden is released under the MIT License. Third-party attribution is listed in THIRD_PARTY_NOTICES.md.
Contributing and security
See CONTRIBUTING.md for development guidelines. Please report security problems according to SECURITY.md, not through a public issue.
For plugin developers
Search results and similarity scores are powered by semantic analysis of your plugin's README. If your plugin isn't appearing for searches you'd expect, try updating your README to clearly describe your plugin's purpose, features, and use cases.