Getting Started

Get ScalyClaw installed and running in under five minutes. This guide covers prerequisites, installation, and your first interaction.

Prerequisites

Dependency	Version	Purpose
Bun	≥ 1.1	JavaScript runtime for all ScalyClaw processes
Redis	≥ 7.0	Message bus, session store, config, vault, pub/sub
Git	any	Cloning the repository

Operating System Support

OS	Status
macOS (Apple Silicon / Intel)	Fully supported
Linux (x64 / arm64)	Fully supported
Windows (via WSL2)	Supported
Docker	Fully supported

Windows Users

Native Windows is not supported. Use WSL2 with a Linux distribution (Ubuntu recommended). Redis and Bun run natively inside WSL.

Installing Bun

bash

curl -fsSL https://bun.sh/install | bash

Installing Redis

bash

# macOS
brew install redis && brew services start redis

# Ubuntu / Debian
sudo apt install redis-server && sudo systemctl enable redis-server

# Docker
docker run -d --name redis -p 6379:6379 redis:7-alpine

Installation

Quick Install (recommended)

bash

curl -fsSL https://scalyclaw.netlify.app/install.sh | bash

This downloads the latest release, installs it to ~/.scalyclaw, and adds the scalyclaw binary to your PATH.

Manual Install (from source)

bash

# Clone the repo
git clone https://github.com/scalyclaw/scalyclaw.git
cd scalyclaw

# Install dependencies
bun install

# Build shared packages
bun run build

First Run

1. Run the Setup Wizard

bash

scalyclaw setup

The wizard walks you through:

Redis connection (defaults to localhost:6379)
Adding your first LLM provider (OpenAI, Anthropic, etc.)
Configuring an embedding model for memory
Setting a dashboard password

2. Start the Processes

bash

# Start the node (brain)
scalyclaw node start

# Start a worker (execution engine)
scalyclaw worker start

# Start the dashboard
scalyclaw dashboard start

Tip

You can run all three in one terminal using scalyclaw start which launches node + worker + dashboard together.

3. Open the Dashboard

Navigate to http://localhost:3000 in your browser. From here you can:

Configure your first messaging channel (e.g., Telegram)
Add LLM models and set priorities
Customize the personality in the Mind page
Send a test message via the Web Gateway

4. Send Your First Message

The fastest way to test is via the Web Gateway. In the dashboard, go to Channels → Web, enable it, then use the built-in chat widget or make an API call:

bash

curl -X POST http://localhost:3000/api/message \
  -H "Content-Type: application/json" \
  -d '{"text": "Hello, ScalyClaw!"}'

Project Structure

scalyclaw/ ├── scalyclaw/ # Node — orchestrator, channels, guards, LLM │ └── src/ │ ├── channels/ # Telegram, Discord, Slack, etc. │ ├── orchestrator/ # Prompt building, LLM routing │ ├── guards/ # Echo, content, skill guards │ └── queues/ # BullMQ queue producers ├── worker/ # Worker — stateless BullMQ consumers │ └── src/ │ ├── processors/ # Message, tool, agent, system processors │ └── executors/ # Code runners (JS, Python, Bash) ├── dashboard/ # Dashboard — React 19 SPA │ └── src/ │ ├── pages/ # admin pages │ └── components/ # Shared UI components ├── shared/ # Shared library — types, utils, prompt, session │ └── src/ │ ├── prompt/ # System prompt sections │ ├── session/ # Per-channel session state machine │ └── memory/ # SQLite + sqlite-vec memory system ├── mind/ # Personality files (user-editable) │ ├── IDENTITY.md # Who it is, tone, boundaries │ ├── SOUL.md # Core values and behavior │ └── USER.md # Baseline user profile ├── skills/ # Skill packages (each in its own folder) ├── agents/ # Agent definitions └── website/ # Landing page and docs