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

Local Agent Assistant

zerozero35 downloads

Ask local AI agents questions about your Markdown vault, cite evidence, and add review notes.

Add to Obsidian
  • Overview
  • Scorecard
  • Updates1

Local Agent Assistant is a desktop Obsidian plugin for asking local AI agent CLIs questions about a Markdown vault, reviewing selected text, and keeping evidence-linked review notes.

Features

  • Select text and open a floating toolbar for explanation, ADR lookup, review suggestions, or comments.
  • Enter a direct instruction without selecting text.
  • Choose among configured local agent CLIs and model identifiers.
  • Search the current Markdown vault before sending bounded evidence to the agent.
  • Keep citations in the form [来源:路径#章节].
  • Add editable AI or manual review comments to the source document.
  • Jump from review locations and evidence references to the source heading.
  • Delete a review by its exact REV-NNN identifier.
  • Interrupt an agent process on Windows, including its child process tree.

Privacy and security

  • No user name, company name, project name, internal URL, account, token, or API key is included in the source or release package.
  • Agent commands are started with an argument array and shell: false.
  • Direct instructions are read-only by default: the plugin asks the agent not to modify files, delete files, or execute shell commands.
  • The plugin does not upload a vault. If the selected agent uses a remote model, the bounded prompt and evidence excerpts may be sent to that model provider.
  • Credentials are managed by the selected agent's own local configuration and are not stored in Markdown, the plugin source, or the release package.

Installation

From the Obsidian Community directory

Install Local Agent Assistant from Settings → Community plugins after the plugin has been published.

From a release ZIP

  1. Extract the release package.
  2. Run install.cmd on Windows.
  3. Choose an Obsidian vault.
  4. Restart Obsidian and enable Local Agent Assistant.

The installer preserves existing community plugins and plugin data. It does not modify Markdown files.

Configure agents and models

Open Settings → Community plugins → Local Agent Assistant.

The Local agent profiles field accepts an array such as:

[
  {
    "id": "hermes",
    "name": "Hermes",
    "command": "hermes",
    "args": ["-z", "{prompt}"],
    "models": ["provider/model-a", "provider/model-b"]
  },
  {
    "id": "my-local-agent",
    "name": "My Local Agent",
    "command": "C:/Tools/my-agent.exe",
    "args": ["--prompt", "{prompt}", "--model", "{model}"],
    "models": ["local-model"]
  }
]

Supported argument placeholders are {prompt}, {vault}, {file}, {selection}, {question}, and {model}. For Hermes, a selected model is passed with -m. For other CLIs, add the model placeholder wherever that CLI expects it.

Release and updates

The plugin is prepared for the Obsidian Community directory:

  • manifest.json uses a unique lowercase hyphenated ID and semantic version.
  • versions.json maps release versions to compatible Obsidian versions.
  • main.js, manifest.json, and styles.css are available as root-level release assets.
  • After publication, each GitHub Release tag must exactly match the version in manifest.json.
  • Increasing the version and publishing a matching GitHub Release makes the update available through Obsidian.

See RELEASE.md for the release checklist.

HealthExcellent
ReviewSatisfactory
About
Ask local AI agent CLIs questions about your Markdown vault or selected text to get explanations, ADR lookups, review suggestions, or comments. Select text to open a floating toolbar, choose a local agent/model, and search the vault for bounded evidence before sending. Attach editable, evidence-linked review notes with citations like [来源:路径#章节], jump to sources, or delete reviews by REV-NNN.
AIReview
Details
Current version
0.4.0
Last updated
Last month
Created
Last month
Updates
1 release
Downloads
35
Compatible with
Obsidian 1.5.0+
Platforms
Desktop only
License
MIT
Report bugRequest featureReport plugin
Author
zerozerofwacjo81
GitHubfwacjo81
  1. Community
  2. Plugins
  3. AI
  4. Local Agent Assistant

Related plugins

LearnKit

A native study system for your Obsidian vault. LearnKit turns notes into durable knowledge with flashcards, spaced repetition, tests, and more.

Claudian

Embeds Claude Code/Codex and other local Agents as AI collaborators in your vault.

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.

Fast Note Sync

Real-time sync of your vaults across server, mobile, and web; shareable with anyone; supports REST and MCP integrations to build your personal AI knowledge base.

ChatGPT MD

A seamless integration of ChatGPT, OpenRouter.ai and local LLMs via Ollama into your notes.

Copilot

Run AI agents such as Claude Code, Codex, and OpenCode inside your vault. Turn your second brain into a smart assistant that gets knowledge work done.

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.

Agent Client

Chat with Claude Code, Codex, Gemini CLI, and more via the Agent Client Protocol — right from your vault.

Text Generator

Generate text content using GPT-3 (OpenAI).

Smart Composer

AI chat with note context, smart writing assistance, and one-click edits for your vault.