Agent Client for Obsidian
Agent Client
An Obsidian plugin that runs Claude Code, Codex and other agents via the Agent Client Protocol right inside your notes vault
High risk
We rate an entry high when the tool writes to external systems, handles money, production databases or secrets, or runs arbitrary commands. The CLI installs it only with your consent.
Why this level
- Launches agents as child processes with full terminal and system access, same as a plain CLI
Install
Manual install
In Obsidian, open Settings, Community Plugins, Browse, search for "Agent Client", click Install, then Enable. Then install and authenticate the agent you need following its own setup guide, and open the chat via the robot icon in the ribbon.
This is third-party code. Review the repository files before installing.
What it does
The plugin adds an AI-agent chat to Obsidian, launched as a separate local process via the Agent Client Protocol (ACP) and working on top of your notes vault. Supported agents include Claude Code, Codex, Gemini CLI, Mistral Vibe, OpenCode, Kiro, Hermes Agent, Pi and Grok Build, and you can wire up a custom agent through the same protocol. The agent reads and writes notes through Obsidian's vault API rather than the filesystem directly, while still getting the same full system access through a terminal as it would in a plain CLI. There's a mode to run the agent inside WSL on Windows.
Who it is for. For Obsidian users who want to work with Claude Code, Codex or another agent without leaving their notes vault.
Good fit when
- You need to discuss and edit notes together with an agent right inside Obsidian
- You already have one of the supported agents installed and want a convenient interface for it in your notes
- You need to run the agent inside WSL from Obsidian on Windows
Not a fit when
- The agent you need doesn't support the Agent Client Protocol and isn't covered by a custom integration
- You're on mobile; the plugin is desktop-only
Example request
Connect Claude Code to Obsidian via Agent Client and help clean up the notes in this sectionLimitations
Desktop-only plugin. Each agent still needs to be installed and authenticated separately, following its own setup guide, before connecting it to Obsidian. The agent gets the same full system access through a terminal as it would running directly on the command line.
How to disable. Open Settings, Community Plugins, and disable or remove Agent Client.
Security check
- Launches agents as child processes with full terminal and system access, same as a plain CLI
README in short
The README describes Agent Client as a way to bring AI agents into Obsidian via the Agent Client Protocol, with ready setup guides for each supported agent: Claude Code, Codex, Gemini CLI, Mistral Vibe, OpenCode, Kiro, Hermes Agent, Pi, Grok Build and custom agents. Install is through Obsidian's built-in community plugin search. A dedicated security section is upfront that the plugin is desktop-only, launches agents as child processes with full terminal access, and that note reading and writing goes through Obsidian's vault API rather than directly touching files.
FAQ
Does the plugin bundle an agent itself?
No, it launches an agent already installed on your machine as a local process via the Agent Client Protocol and gives it an interface inside Obsidian.
What access does the agent get to notes?
Reading and writing notes goes through Obsidian's vault API, but the agent process itself keeps ordinary terminal and system access, same as running it directly.
Related
A self-hosted knowledge base with block-level references and a built-in MCP server for connecting AI agents to your notes
A CLI for every Google Workspace API with JSON output and agent skills: Drive, Gmail, Calendar, Sheets and more
Local search over Markdown notes, docs and meeting transcripts: keywords, semantic search and reranking, with an MCP server
A task manager for AI-driven development: breaks a PRD into dependent tasks and guides the agent through them via MCP or CLI