FastClaw / Get started

quick start

From installing a single binary to entering the Dashboard: run it first, then create the provider, agent and first session.

View official source Docs index

Installation

The official README provides a one-line installation script. After the installation is complete, running fastclaw directly will enter the first startup process and open the Dashboard on the local machine.

curl -fsSL https://raw.githubusercontent.com/fastclaw-ai/fastclaw/main/install.sh | bash
fastclaw

first start

The first startup will create FastClaw home, database and default configuration, and guide you to set up LLM provider and default Agent. The default Dashboard address is http://localhost:18953.

Start gateway

Running fastclaw, the process will load env, install bundled skills, start API server and Dashboard.

Complete setup

After the browser opens the local page, create an administrator or use a setup token to complete the initialization.

Configure provider

Add model providers such as OpenAI, Anthropic, Ollama, OpenRouter and more in Dashboard or CLI.

Create Agent

Select the model, write SOUL/IDENTITY, confirm the tools and sandbox, and start your first session.

First conversation

It is recommended to use low-risk agents to verify models, tools, and file permissions first, rather than connecting them to the production workflow.

fastclaw agents init demo \
  --provider openai \
  --model openai/gpt-4o-mini \
  --api-key-env OPENAI_API_KEY

fastclaw agents ls

Start checking

After running through, confirm four things: the Dashboard can be logged in, the provider can complete a conversation, the Agent file can be read and written, and the sandbox tool is limited as expected.

Dashboard: http://localhost:18953
Default port: FASTCLAW_PORT, default 18953
Default data directory: ~/.fastclaw
Running in the background: fastclaw daemon start