Skip to main content

Getting Started

This guide will help you get Ragpi up and running in minutes. We'll walk through deploying Ragpi locally, adding your first documentation source, and making your first query.

Prerequisites​

  • Docker and Docker Compose installed
  • Git installed
  • An OpenAI API key

Deploying Ragpi Locally​

  1. Clone the repository and navigate to the directory:
git clone https://github.com/ragpi/ragpi.git
cd ragpi
  1. Copy the example environment file and open it for editing:
cp .env.example .env
  1. Configure the essential environment variables in .env:
# Add your OpenAI API key
OPENAI_API_KEY=your_api_key_here

# Optional: Add API authentication
API_KEYS=your_secret_api_key
info

If you would like to enable API authentication, add a comma-separated list of keys to the API_KEYS variable. You will need to generate these keys yourself and include them in the x-api-key header for all requests.

  1. Start Ragpi using Docker Compose:
docker compose -f docker-compose.prod.yml up -d

Your Ragpi instance is now running at http://localhost:8000! You can verify it's working by visiting the OpenAPI documentation at http://localhost:8000/docs.

Adding Your First Source​

Let's add a documentation website as our first source. We'll use the Sitemap connector to automatically process all pages from a sitemap. You can replace the parameters in the example below with your own values.

  1. Create a source using the /sources endpoint:
curl -X POST http://localhost:8000/sources \
  -H "Content-Type: application/json" \
  -H "x-api-key: your_secret_api_key" \
  -d '{
    "name": "example-docs",
    "description": "Documentation for example project",
    "connector": {
      "type": "sitemap",
      "sitemap_url": "https://docs.example.com/sitemap.xml"
    }
  }'
  1. Note the task_id in the response. You can monitor the synchronization progress:
curl http://localhost:8000/tasks/{task_id} \
  -H "x-api-key: your_secret_api_key"

Making Your First Query​

Once your source has finished synchronizing, you can start asking questions! Let's try a simple query:

curl -X POST http://localhost:8000/chat \
  -H "Content-Type: application/json" \
  -H "x-api-key: your_secret_api_key" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "What are the deployment options for this project?"
      }
    ]
  }'

The response will include Ragpi's answer, synthesized from the relevant documentation it found. Since no sources are specified, Ragpi will query all available sources. If you want to limit the search to a specific source, you can include the source field in the request body.

Next Steps​