Search...Search plugins and themes...
⌘K
Sign in
  • Get started
  • Download
  • Pricing
  • Enterprise
  • Account
  • Obsidian
  • Overview
  • Sync
  • Publish
  • Canvas
  • Mobile
  • Web Clipper
  • CLI
  • Learn
  • Help
  • Developers
  • Changelog
  • About
  • Roadmap
  • Blog
  • Resources
  • System status
  • License overview
  • Terms of service
  • Privacy policy
  • Security
  • Community
  • Plugins
  • Themes
  • Discord
  • Forum / 中文论坛
  • Merch store
  • Brand guidelines
Follow us
DiscordTwitterBlueskyThreadsMastodonYouTubeGitHub
© 2026 Obsidian

Analogy - RAG in your vault

liaocaoxuezheliaocaoxuezhe1k downloads

Search your vault semantically with local vector embeddings and expose RAG search to MCP clients.

Add to Obsidian
  • Overview
  • Scorecard
  • Updates14

Analogy adds local semantic search to Obsidian and exposes your indexed vault to MCP clients for retrieval-augmented generation.

The plugin is desktop-only because it runs a managed local runtime (Node.js, ONNX Runtime) for embeddings and a local SQLite vector store.

Features

  • Index Markdown notes into local vector chunks.
  • Search the current vault semantically from the Analogy side pane.
  • Explore related indexed chunks in a local Semantic Walk canvas.
  • Exclude folders or files from the local index.
  • Hybrid search: semantic similarity fused with keyword matching for exact names, people and code identifiers.
  • Built-in MCP server with one-click setup for Claude Desktop, Claude Code, Cursor, opencode and Codex.

Install

Community plugin release

Download the release assets from GitHub and copy them into:

<Vault>/.obsidian/plugins/analogy-rag-in-your-vault/

Required release assets:

  • main.js
  • manifest.json
  • styles.css

Enable the plugin in Obsidian Settings -> Community plugins.

Local runtime

Obsidian installs only the three files above; Analogy's guided onboarding sets up the rest. It downloads the managed embedding runtime for your platform from this repository's runtime-<version> release and one embedding model, verifies every file against hashes pinned in the plugin, and builds a first index. No Python, npm install or command line is needed. Supported: macOS on Apple Silicon and Windows x64.

  • Each platform has its own runtime version, so a Windows runtime update never re-downloads anything on macOS. When a runtime is superseded, Analogy offers an Upgrade the inference engine step that keeps your index and models.
  • Downloads resume after interruptions and follow the system proxy. Without network, use Settings → Local runtime → Offline install to import the runtime archive and model folder; both are verified before use.
  • The model picker shows download size and estimated memory; first installs choose a multilingual model for Chinese-first vaults.

Details: docs/local-runtime.md. Error codes: docs/install-troubleshooting.md.

Usage

  1. Open Obsidian and enable Analogy.
  2. Click the Analogy ribbon icon to open the side pane.
  3. Open Settings -> Analogy - RAG in your vault.
  4. Verify that the local runtime and the embedding model show as ready.
  5. Click Continue index or Rebuild index.
  6. Search your vault from the Analogy pane, or connect an AI client under Settings → MCP (docs/mcp-server.md).

Semantic Walk

Semantic Walk is a local, chunk-level exploration canvas in the main Obsidian workspace. Open it from the ribbon, the command palette, the current document, a sidebar search result, or a random indexed chunk.

Expanding a node queries the existing local index on demand. Multiple chunks from the same Markdown document can remain on the canvas while you follow another branch. Use New batch to reset the graph and choose a new random root. Relationship labels are relative to the current candidate batch rather than an absolute similarity percentage.

Indexes created by older Analogy versions may lack stable chunk identity, heading, or position metadata. Existing search remains compatible, but run Rebuild index to enable every Semantic Walk entry and obtain accurate chunk positioning.

Privacy and data flow

  • Your vector index is stored in a device-local data folder outside the vault (~/Library/Application Support/Analogy or %LOCALAPPDATA%\Analogy) and is never synced.
  • Semantic Walk reuses the local chunk index and embedding worker; it does not create a remote graph service or upload note content.
  • The vector store is a local SQLite file; no network service is opened.
  • Markdown content is read from the current vault for indexing.
  • Embedding model files may be downloaded from the configured model host, which defaults to https://hf-mirror.com/.
  • License validation uses https://analogy.zexing.club/api/v1/obsidian/license/validate; deactivation uses /api/v1/obsidian/license/deactivate. These requests send the license key, plugin version, a local device identifier, and a vault identifier, but never note contents, paths, queries, or embeddings.
  • Purchase and account-management actions open https://analogy.zexing.club/analogy and https://analogy.zexing.club/analogy/account in the browser. The plugin does not receive Stripe credentials or payment-card data.
  • The plugin caches license state locally and normally refreshes it about every 7 days. If the service is temporarily unavailable, the last active state remains usable only through the server-provided grace_until; after that, the plugin falls back to the free limits without blocking local note access.
  • MCP clients can access search snippets only when you explicitly configure them. Writing a client configuration shows the exact change first and keeps a backup. The MCP server records only the time, tool and result count of the last query, never the query text.
  • Use ANALOGY_ALLOWED_PATHS to restrict which vault paths the MCP server can return.
  • Semantic Walk diagnostics contain only a hashed chunk identifier and allowlisted runtime metadata; they exclude chunk text, paths, queries, embeddings, and license keys.

Development

npm install
npm run dev

Build for CI without deploying to Obsidian or creating release assets:

npm run build:ci

Build and atomically deploy the three-file plugin to the configured local Obsidian vault:

npm run build:local

Prepare release assets:

npm run release:prepare

The release tag must exactly match manifest.json version.

Repository layout

  • src/ - Obsidian plugin source.
  • src/mcp/server/ - MCP server bundled into main.js; reuses the read-only store reader and config in mcp-server/src/.
  • runtime-package/ - Reviewed managed-runtime inputs, including the per-platform version table.
  • scripts/ - Runtime build, verification and release scripts.
  • benchmark/ - Hybrid retrieval and vector scan benchmarks.
  • .obsidian/ - Local development vault, ignored by the standalone plugin repo.

License

Apache-2.0

HealthExcellent
ReviewRisks
About
Index your vault and retrieve notes relevant to queries to build a local knowledge base. Generate context-aware answers or concise summaries with a connected language model that reference your notes.
AISearchResearch
Details
Current version
1.3.2
Last updated
2 days ago
Created
5 months ago
Updates
14 releases
Downloads
1k
Compatible with
Obsidian 1.12.7+
Platforms
Desktop only
License
Apache-2.0
Report bugRequest featureReport plugin
Sponsor
Support
Author
liaocaoxuezheliaocaoxuezheliaocaoxuezhe
liaocaoxuezhe.com/
GitHubliaocaoxuezhe
Xliaocaoxuezhe
  1. Community
  2. Plugins
  3. AI
  4. Analogy - RAG in your vault

Related plugins

Local LLM Helper

Use a local LLM server to augment your notes

Vault Curate

A local-first second brain for Obsidian , Desktop / Mobile ok, semantic search · relation graph · semantic paths · Hot/Cold rediscovery · strong Chinese/CJK · WebGPU on-device · no API keys. RAG LLM.

Smart Lookup

Semantic search for your vault. Ask in natural language, find notes by meaning when exact words fail, preview matching notes, and turn forgotten ideas into links, context, and next steps.

Smart Connections

Find related notes and excerpts while writing. Your AI link building copilot displays relevant content in graph + list view. A local embedding model powers semantic search. Zero setup. No API key.

Karpathy LLM Wiki

Karpathy's LLM Wiki implementation plugin for Obsidian - turns notes and PDFs into a linked, LLM-powered knowledge base with entity pages, concept pages, graph-powered Q&A, and local-first privacy.

Smart Second Brain

Chat with an agent that knows your notes, search by meaning, and see your vault grouped into topics. Runs on Ollama, OpenAI, Anthropic, and more.

Khoj

An AI personal assistant for your digital brain.

Citations

Automatically search and insert citations from a Zotero library.

AI image analyzer

Analyze images with AI to get keywords of the image.

Similarity

Search, explore and see related notes based on their meaning (semantics). Using a tiny AI running fully on-device: data never leaves your machine, no API keys needed. Especially great for journaling.