Skip to main content
This guide will help you go from zero to a working CodeFire setup with persistent memory for your AI coding agent.
1

Download and Install

Download CodeFire.app from GitHub Releases and drag it to your Applications folder.
On first launch, macOS may show a security warning. Right-click the app and select “Open” to bypass Gatekeeper.
2

Connect Your AI CLI Tool

Configure the MCP server for your preferred AI coding tool. You can use the one-click installer in the app, or configure manually:
Run this command in your terminal:
claude mcp add codefire ~/Library/Application\ Support/CodeFire/bin/CodeFireMCP
Add this to ~/.gemini/settings.json:
{
  "mcpServers": {
    "codefire": {
      "command": "~/Library/Application Support/CodeFire/bin/CodeFireMCP",
      "args": []
    }
  }
}
Add this to ~/.codex/config.toml:
[mcp_servers.codefire]
command = "~/Library/Application Support/CodeFire/bin/CodeFireMCP"
args = []
Add this to ~/.opencode/config.json:
{
  "mcpServers": {
    "codefire": {
      "command": "~/Library/Application Support/CodeFire/bin/CodeFireMCP",
      "args": []
    }
  }
}
The MCP server binary is automatically installed to ~/Library/Application Support/CodeFire/bin/CodeFireMCP when you first launch the app.
3

Configure OpenRouter API Key

Open CodeFire Settings (Cmd + ,), navigate to the CodeFire Engine tab, and paste your OpenRouter API key.This powers:
  • Built-in AI chat with full project context
  • Semantic code search
  • AI image generation
All models are served through OpenRouter with a single key.
Without an OpenRouter API key, AI-powered features like chat and semantic search will not work.
4

Verify the Connection

Test that everything is working:
  1. Launch your AI CLI tool in a project directory: 
    claude code
# or: gemini cli
# or: codex
# or: opencode
  2. Ask your AI agent to list tasks: 
    Can you show me my current tasks?
  3. Check CodeFire — you should see:
    • Your project auto-discovered in the sidebar
    • Live session monitoring showing the active conversation
    • Token usage and cost tracking
If the MCP connection is working, your agent will be able to read and create tasks, access notes, and query your project structure.
5

Start Coding with Persistent Memory

Your AI agent now has:
  • Access to all tasks and project notes
  • Awareness of your project structure and session history
  • The ability to create and update tasks programmatically
Try asking your agent to:
  • “Create a task to refactor the authentication module”
  • “What notes do I have for this project?”
  • “Show me the project file tree”
All interactions are tracked in CodeFire, creating a persistent memory layer across sessions.

​
Optional Setup

To enable automatic task creation from emails:
  1. Go to Settings → Gmail
  2. Enter your Google OAuth credentials (Client ID + Client Secret)
  3. Get these from Google Cloud Console → APIs & Services → Credentials
Once configured, CodeFire can extract tasks from your emails automatically.
Install GitHub CLI to enable:
  • PR viewing and status
  • CI/CD status monitoring
  • Issue tracking
brew install gh
gh auth login

​
Next Steps

Installation Details

Learn about system requirements, troubleshooting, and advanced setup

Features

Explore all CodeFire features in depth

Build docs developers (and LLMs) love

Get started for freeTalk to us