Getting Started

Download CrabDash, add your first provider key, and send a request through the local gateway.

CrabDash is a macOS menubar app that gives you a personal LLM gateway. Every request — cloud or local — routes through one port on your machine. Nothing leaves localhost unless you tell it to.

Requirements

macOS 14 (Sonoma) or later
Apple Silicon (M1+) for local models, Intel supported for cloud-only

Install

Download CrabDash.dmg. Open the .dmg, drag to Applications, and launch. The crab icon appears in your menubar.

First launch

On first launch CrabDash starts a local CrabLLM gateway on 127.0.0.1:5635. It speaks both OpenAI and Anthropic formats — point any compatible app or SDK at it.

curl http://127.0.0.1:5635/v1/models

No models will appear yet. You need at least one provider key.

Add a provider key

Click the menubar icon
Open Settings → Providers
Click Add Provider, pick a provider (e.g., OpenAI), and paste your API key
The models from that provider appear immediately in the request dropdown and in /v1/models

See Provider Keys for step-by-step instructions on getting free keys from each provider.

Send your first request

curl http://127.0.0.1:5635/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [{"role": "user", "content": "Hello from CrabDash"}]
  }'

The request appears in the menubar's live tail. You can see the provider it routed to, latency, token count, and cost.

Next steps

Get free provider keys for Google Gemini, Groq, NVIDIA, and more
Use the gateway API from any app or SDK
Run local models via MLX on Apple Silicon