Vault Chat Agent

approved

by Evgenii Bykovskikh

Chat with an AI assistant that can search your vault, inspect notes, and prepare edits for review - This plugin has not been manually reviewed by Obsidian staff.

↓ 277 downloadsUpdated 1mo agoMIT
Install in ObsidianView on GitHub

Vault Chat Agent

AI assistant plugin for working with your Obsidian vault.

What Works

  • One agentic chat with Ask and Edit intents.
  • OpenAI-compatible chat completions via /v1/chat/completions.
  • Search scopes for whole-vault, current-note, and current-folder chat.
  • Plan mode for Edit chats, allowing read-only investigation before pending edits are prepared.
  • @current, @[[Note]], @path.md, and @folder/ context mentions with inline autocomplete.
  • Provider presets for OpenAI, DeepSeek, OpenRouter, LM Studio, Ollama, and custom endpoints.
  • Model browsing from OpenAI-compatible /v1/models endpoints.
  • Local vault indexing across supported text files, Canvas files, and metadata-only attachments.
  • Graph-aware retrieval using note links, backlinks, and neighboring chunks.
  • Related Notes sidebar based on the active note or selected text.
  • Read-only agent tools for searching, opening notes, inspecting links, and reviewing the current note, with clickable source references.
  • Text commands for summarizing selections, extracting action items, and proposing polished rewrites.
  • Saved prompts with Ask/Edit intent, runnable from the command palette against the current selection or note.
  • Manual workflows for weekly reviews, meeting tasks, and project status notes, all behind pending edit review.
  • Reviewed edit proposals with patch previews, Apply/Reject, and batch patch support.
  • Optional stripping of <think>, <reasoning>, and <thought> blocks from assistant output.
  • Realtime re-indexing on vault create, modify, delete, and rename events.

Build

npm install
npm run build

Then copy the contents of dist/ into:

<vault>/.obsidian/plugins/vault-chat-agent/

Enable the plugin in Obsidian community plugin settings, then configure an OpenAI-compatible provider in the plugin settings.

Provider Notes

The plugin calls:

<API base URL>/v1/chat/completions

Remote providers usually need an API key. Local providers such as LM Studio and Ollama can leave the API key empty when using a localhost base URL.

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.