Quick Start - SpillwaveSolutions/agent-brain GitHub Wiki

Agent Brain Quick Start

Get up and running with Agent Brain in minutes. The Claude Code plugin is the primary interface - once installed, it handles everything else for you.

Step 1: Install the Plugin

Install the Agent Brain plugin in Claude Code:

claude plugins install github:SpillwaveSolutions/agent-brain

This gives you access to 24 commands, 3 intelligent agents, and 2 skills for working with Agent Brain.

Step 2: Install the Server and CLI

Use the plugin to install the Python packages:

/agent-brain-install

This installs:

  • agent-brain-rag - The FastAPI server for indexing and search
  • agent-brain-cli - Command-line tool (used internally by the plugin)

Step 3: Configure API Keys

Configure your embedding and summarization providers:

/agent-brain-providers

Choose from:

  • Cloud providers: OpenAI, Anthropic, Cohere, Gemini, Grok
  • Local providers: Ollama (fully offline mode)

Or use the complete setup wizard which handles installation AND configuration:

/agent-brain-setup

Step 4: Initialize Your Project

Initialize Agent Brain for your current project:

/agent-brain-init

This creates a .claude/doc-serve/ directory with project-specific configuration.

Step 5: Start the Server

Start the Agent Brain server:

/agent-brain-start

The server starts with automatic port allocation (no conflicts with other projects).

Step 6: Index Your Documentation

Index your project's documentation and code:

/agent-brain-index ./docs

For code + documentation:

/agent-brain-index . --include-code

Check indexing status:

/agent-brain-status

Step 7: Search Your Knowledge Base

Now you can search! Use the smart search command:

/agent-brain-search "how does authentication work"

Or use specific search modes:

/agent-brain-semantic "explain the architecture"
/agent-brain-keyword "getUserById"
/agent-brain-graph "what calls AuthService"

All-in-One Setup

For the fastest setup, use the interactive wizard which does steps 2-6 automatically:

/agent-brain-setup

The Setup Assistant guides you through:

  1. Installing packages
  2. Configuring API keys
  3. Initializing the project
  4. Starting the server
  5. Indexing your documentation

Search Modes Quick Reference

Command Best For Example
/agent-brain-search General questions "how does caching work"
/agent-brain-semantic Conceptual queries "explain the data flow"
/agent-brain-keyword Exact terms, errors "NullPointerException"
/agent-brain-hybrid Fine-tuned search "API authentication"
/agent-brain-graph Dependencies "what uses UserService"
/agent-brain-multi Maximum recall "everything about validation"

Using Agents for Complex Tasks

For complex research tasks, Agent Brain's intelligent agents help:

You: "Research how error handling is implemented across the codebase"

Research Assistant automatically:

  1. Searches documentation for error handling patterns
  2. Queries code for try/catch blocks and error classes
  3. Uses graph mode to find error propagation
  4. Synthesizes a comprehensive answer with file references

Verify Your Setup

Check that everything is working:

/agent-brain-verify

This validates:

  • Package installation
  • API key configuration
  • Server connectivity
  • Index health

Next Steps