Skip to main content
This guide walks you through the complete installation process for Athena, from prerequisites to your first working session.

Prerequisites

Before you begin, ensure you have:

Step 1: Clone the repository

Clone the Athena repository to your local machine:
git clone https://github.com/winstonkoh87/Athena-Public.git
cd Athena-Public
Clone it anywhere you keep projects (e.g. ~/Projects/). This folder is your Athena workspace — your memory, protocols, and config all live here.

Step 2: Set up virtual environment

On macOS (Homebrew) and Ubuntu 23.04+, installing packages without a virtual environment will fail with externally-managed-environment. Always create a .venv first.
Create and activate a Python virtual environment:
python3 -m venv .venv
source .venv/bin/activate
You should see (.venv) appear in your terminal prompt, indicating the virtual environment is active.

Step 3: Install the SDK

Athena offers two installation modes:
# ~30 seconds, no ML dependencies
pip install -e .
The lightweight install is perfect for getting started. You can always upgrade to the full install later by running pip install -e ".[full]".
Don’t pip install athena-cli — that’s a different, unrelated package. Always install from inside the cloned repo using pip install -e .

Step 4: Initialize workspace (optional)

If you want to create a separate Athena workspace outside the cloned repo:
# Navigate to your desired location
cd ~/MyAssistant

# Initialize Athena workspace
athena init .
Most users work directly inside the cloned Athena-Public/ folder. Only create a separate workspace if you want to keep your personal data isolated from the example files.
This creates the full structure:
MyAssistant/
├── .athena_root          # Workspace marker
├── .agent/
│   ├── workflows/        # start.md, end.md, save.md
│   ├── scripts/
│   └── skills/protocols/
├── .framework/modules/   # Core_Identity.md
└── .context/
    ├── project_state.md
    └── memories/session_logs/

IDE-specific initialization

You can generate IDE-specific configuration files during initialization:
athena init . --ide cursor

Step 5: Configure your IDE

Open the Athena-Public/ directory (or your custom workspace) as your workspace root in your AI-enabled IDE.

Cursor

  1. Open the Athena folder: File → Open Folder
  2. Cursor will automatically detect .cursor/rules.md if you ran athena init --ide cursor
  3. Press Cmd+L (macOS) or Ctrl+L (Windows/Linux) to open the AI chat panel

Antigravity

  1. Open the Athena folder in your browser via Antigravity
  2. Antigravity will automatically load AGENTS.md if present
  3. Use the chat sidebar to interact with your agent

VS Code + Copilot

  1. Open the Athena folder: File → Open Folder
  2. Install the GitHub Copilot extension if not already installed
  3. Open Copilot Chat: Ctrl+Alt+I (Windows/Linux) or Cmd+Shift+I (macOS)

Claude Code

  1. Launch Claude Code from your terminal: claude-code
  2. Open the Athena folder
  3. Use the integrated chat panel
Athena does NOT work through ChatGPT.com, Claude.ai, or Gemini web. You need an app that can read files from your disk.

Step 6: Verify installation

Run the system health check:
athena check
You should see output like:
🩺 ATHENA SYSTEM CHECK
============================================================
   CLI Version: 9.3.1
   ✅ Workspace marker: Found
   ✅ .agent/: Found
   ✅ .context/: Found
   ✅ .framework/: Found
============================================================

Step 7: Boot your first session

In your IDE’s AI chat panel (not the terminal), type:
/start
/start is an AI chat command — you type it in the chat window where you talk to the AI, not in the terminal.
You should see your agent load its core identity and confirm it’s ready.

Next steps

Your first session

Walk through a guided tutorial

Architecture

Understand how Athena works

Troubleshooting

”No session log found”

Run /start first to create a session log before using other commands.

Scripts don’t run

Check that Python is installed and accessible:
python3 --version
Should output Python 3.10 or higher.

Workflow not recognized

Ensure workflow files are in .agent/workflows/ and your IDE can access them.

Windows encoding errors

$env:PYTHONIOENCODING="utf-8"
Or use Windows Terminal from the Microsoft Store.

Permission denied (macOS/Linux)

Make scripts executable:
chmod +x .agent/scripts/*.py

Virtual environment not activating

# Try with explicit path
source .venv/bin/activate

# Or use python -m
python3 -m venv .venv --clear
source .venv/bin/activate

Advanced configuration

Custom workspace location

Work in a different directory:
athena init ~/Documents/MyAthena
cd ~/Documents/MyAthena

Multi-root workspace

Work on external projects while keeping Athena’s memory:
  1. Open Athena folder in your IDE
  2. File → Add Folder to Workspace
  3. Add your project folder
Now you can work on your project while Athena maintains context across both folders.

Environment variables (optional)

For cloud features (Supabase sync, API integrations), create a .env file:
# .env
SUPABASE_URL=your_supabase_url
SUPABASE_ANON_KEY=your_anon_key
ANTHROPIC_API_KEY=your_api_key  # if using Claude directly
Cloud features are optional. Athena works 100% locally by default.

Build docs developers (and LLMs) love