Skip to content
FoxwarmFoxwarmFoxwarm

Quick start

This page takes the shortest path to a working agent. Each step links to the full guide if you need more options.

1. Install

Section titled “1. Install”

Run the command for your platform from the directory where you want Foxwarm’s program and data directories created.

Linux x64, x64 WSL or macOS on Apple Silicon

Terminal window
curl -fsSL https://foxwarm.net/install-foxwarm.sh | bash

Windows x64 (PowerShell)

Terminal window
irm https://foxwarm.net/install-foxwarm.ps1 | iex

Docker (Linux x64)

Terminal window
mkdir -p foxwarm-data
docker run -d --name foxwarm --init -p 3001:3001 \
  -v "$PWD/foxwarm-data:/data" ghcr.io/550w-host/foxwarm:latest

The installers download a prebuilt release; nothing is compiled. When the ASP.NET Core 10 runtime is not installed, they use the package that includes .NET. On Windows there is also a desktop app. See Install Foxwarm for packages and requirements.

When the installer finishes, it starts Foxwarm and prints a sign-in link such as http://localhost:3001/#token=.... Treat that link like a password.

2. Add a model

Section titled “2. Add a model”

Open the sign-in link. On a new installation, Foxwarm opens Setup → Models. Choose your service, paste your API key, pick a model (Load models lists what the service offers) and select Test connection. Then save and start chatting.

Prefer YAML? Switch the Models view to YAML and paste a provider configuration, for example:

default: my-provider/my-model
providers:
  my-provider:
    providerType: openai-completions   # or openai-responses, anthropic
    baseUrl: https://api.example.com/v1
    apiKey: replace-with-your-key
    models:
      - my-model

Set up your first model explains each field.

3. Give an agent a task

Section titled “3. Give an agent a task”

After the first save, Foxwarm creates the main Agent’s default Session main with a short welcome that explains Agents and Sessions and asks what to call you. Closing Setup takes you there. Reply to it: your answer is also the first live test of the provider connection, and the Agent remembers what you tell it. Then give it something real, in a directory you are happy for it to work in:

Look at ~/projects/my-app. Run the tests and tell me which ones fail and why. Don't change anything yet.

You will see the agent’s tool calls and results in the conversation. The Session keeps its history, so you can come back later and continue where you left off, even after a restart.

Where to go next

Section titled “Where to go next”