๐Getting Started with Gclaw
Install Gclaw, run onboarding, and launch your first living AI trading agent
Living Agent features are enabled out of the box โ no extra configuration needed. Run gclaw onboard, follow the prompts, and your agent starts its metabolism immediately.
This guide walks you through everything you need to get Gclaw running: installing the binary, completing onboarding, launching the agent, and verifying it's alive via the Living Dashboard.
Prerequisites
Before you begin, make sure you have:
Go 1.21 or later (if installing from source) โ download at go.dev/dl
A GDEX API key (optional but recommended) โ enables live trading via GDex Pro; without it, Gclaw runs in market-monitor mode only
50 MB of free disk space โ the Gclaw binary is under 10 MB; the rest is config and agent memory
Gclaw is designed to run on minimal hardware. It will work on a Raspberry Pi, a $10 IoT board, a cheap VPS, or your laptop.
Step 1: Install Gclaw
Option A โ Install with Go:
go install github.com/GemachDAO/Gclaw/cmd/gclaw@latestOption B โ Download a pre-built binary:
Download the latest release for your platform from the Gclaw releases page, then move it to your PATH:
# Example for Linux/macOS (amd64)
chmod +x gclaw
mv gclaw /usr/local/bin/gclawVerify the installation:
Step 2: Run Onboarding
Onboarding configures your LLM provider, optional GDEX trading credentials, and initial agent parameters.
The interactive onboarding flow will ask you to:
Choose your LLM provider โ OpenAI, Anthropic, or Ollama (local)
Enter your API key โ stored locally in
~/.gclaw/config.jsonConfigure GDEX trading (optional) โ enter your GDEX API key and wallet address
Set your risk preferences โ initial position size, maximum drawdown, and preferred chains
Step 3: Start the Agent
Once onboarding is complete, launch your agent:
This starts the Living Agent process. You will see the agent's startup sequence in the terminal โ including its initial GMAC balance, the LLM provider it connected to, and confirmation that the metabolism has started ticking.
What Happens Automatically
As soon as the agent starts, several systems activate without any additional setup:
GMAC metabolism begins โ the agent starts at 1,000 GMAC and begins burning a small amount per heartbeat tick
Market scanning starts โ Gclaw immediately begins monitoring token prices and trends via GDEX
GDEX trading is enabled (if credentials were provided) โ the agent evaluates trade opportunities based on its default strategy
Skills load โ all built-in skills (trading, portfolio management, token discovery, etc.) are available immediately
Living Dashboard launches โ accessible at
http://127.0.0.1:18790
The Living Dashboard
Open your browser and navigate to:
The dashboard gives you a real-time view of your agent's life state:
GMAC balance โ current token reserves and burn rate
Goodwill score โ reputation progress toward the next ability threshold
Active trades โ open positions and recent trade history
Heartbeat status โ tick frequency and inference cost log
Family tree โ child agents (if any have been spawned via self-replication)
Configuration
Gclaw stores its configuration at ~/.gclaw/config.json. You can edit this file directly to customize metabolism settings, trading parameters, and agent behavior.
Example metabolism configuration:
Changes to the config file take effect on the next agent restart. Agents with goodwill โฅ 100 can also self-modify some of these parameters autonomously โ see ๐งฌ The Living Agent for details.
Next Steps
Your agent is live. Here's what to explore next:
๐งฌ The Living Agent โ Understand GMAC metabolism, goodwill thresholds, and survival mechanics
๐ง Agent Skills โ Browse and install additional skill modules
๐ Trading with Gclaw โ Learn how Gclaw trades across chains and how to tune its strategy
๐ Swarm & Replication โ Once you hit goodwill โฅ 50, spawn your first child agent
Last updated