Quickstart

From nothing to your first answer in about a minute.

1. Install

curl -fsSL https://sportsclaw.gg/install.sh | bash

This installs the sportsclaw command, the sports data layer, and a process manager for running bots in the background. You'll need Node.js 18+ and Python 3.9+ already on your machine — the installer checks for both and tells you how to get them if they're missing.

Prefer to install manually?

npm install -g sportsclaw-engine-core pm2
python3 -m pip install --upgrade sports-skills

2. Connect a model

sportsclaw uses your AI model to do the reasoning — the sports data itself is free and needs no key. Pick one:

# Use an API key from Anthropic, OpenAI, Google, or Azure Foundry:
export ANTHROPIC_API_KEY=sk-...

# …or, if you already use Claude Code, reuse that session:
sportsclaw login claude

Prefer a guided setup? Run sportsclaw config and it'll walk you through choosing a provider and model. See Configuration for all the options.

3. Ask

sportsclaw "What are today's NFL scores?"

That's it. The agent figures out which data it needs, fetches it live, and answers:

❯ sportsclaw "What are today's NFL scores?"

  ◒  Thinking…
  ✓  fetched NFL scores

  Here are today's NFL results:
  • Chiefs 27, Raiders 20 (Final)
  • Eagles 31, Cowboys 17 (Final)
  • Lions at Packers — 8:20pm ET

The first time you ask about a sport, sportsclaw installs it on the spot (a couple of seconds). To pre-install everything up front, run sportsclaw init --all.

Keep going

sportsclaw chat

chat opens an interactive session — ask follow-ups, dig into odds, compare teams.

What's next

• Configuration — choose your model, manage providers, set up bots
• How It Works — what the engine is doing under the hood
• Building a Discord bot — put it in front of your community