User guide¶
You want to use Particles to manage knowledge — deposit sources, extract claims, query, export.
This guide is task-oriented: each page walks you through one job end-to-end. Read Getting started first; the other pages are reference-style and you can dip into whichever you need.
| Page | When to read |
|---|---|
| Getting started | First-time setup → first query |
| Concepts | What's a particle, subject, status, confidence, provenance |
| Querying | Semantic search, tag filters, structural filters, MCP, what the ranking does |
| As-of time travel | What the store believed at a past instant, and what replaced it |
| Exporting | Obsidian vault, Anki deck, wiki articles, Logseq graph |
| Graph view | The scoped epistemic subgraph as a self-contained HTML file |
| Using from LangChain | Consume a Particles store as LangChain tools / a retriever |
| Claude Code memory | Wire a store in as managed agent memory — session-start push, session-end harvest |
| Swapping in for the reference memory server | Drop-in replacement for @modelcontextprotocol/server-memory |
| Depositing from your phone | Deposit URLs from the iOS Share Sheet via HTTP or an iCloud inbox file |
For operator-side concerns (long-term tuning, lint hygiene, troubleshooting), see the operator guide. For writing a new extractor or exporter, see the plugin-author guide.