Caprock — Claude Code MCP server: what it does, how to install
Caprock — an MCP server for Claude Code, Codex, and other agents that visualizes sessions, history, and API costs. Tracks task execution, CI status, and usage limits in real time.
Apache-2.0 ★ 50 updated 2026-11-16
Open on GitHubWhat Caprock does
Caprock is a desktop application for macOS, Windows, and Linux that centralizes data from your coding agents into one interface. It displays active sessions, allows you to approve agent actions via notifications, and preserves chat history even after the agent clears its local context.
The tool monitors API expenses with limit warnings and automates pull request management, including conflict resolution and branch updates. It also supports Obsidian vault integration and provides an MCP server to let agents access your project data and history.
What it helps with
- track API costs in real time
- approve agent actions via notifications
- preserve session history after context clearing
- fix CI errors and conflicts using agents
- integrate Obsidian notes into agent context
Example prompts for the agent
- Use
caprock mcpto connect your agent to the session data. - Approve or reject agent commands directly from the Caprock notification.
- View pull request progress and CI status within the app interface.
How to install Caprock
Install
The MCP registry has no launch command for it: see the repository README
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Caprock
How to install Caprock?
Install it on macOS, Windows, or Linux using brew install --cask dspv/tap/caprock-app or download the installer from the GitHub releases page.
What does Caprock do?
It is a monitoring app for coding agents that aggregates terminal sessions, request history, API limit tracking, and pull request management into a single GUI.
What is required to run Caprock?
The app runs locally with no account required. To enable agent integration, simply configure the connection using caprock mcp within your project.