Skip to main content

Playground

The Playground provides a unified interface to test both semantic search and AI-powered chat against your indexed documents. It features two tabs: Search for testing vector similarity queries, and Chat for conversational RAG (Retrieval-Augmented Generation) interactions.

All authenticated users have access to the Playground—no special permissions required.

Prerequisites​

Before using the Playground, ensure you have:

  • Vector store configured: Set up in Settings > Vector Store
  • Embedding provider set up: Set up in Settings > Embedding Model. Cloud providers need their API key saved; an OpenAI key alone is enough, because the default model is OpenAI text-embedding-3-small, while other providers also need Save Settings. Ollama needs no key: save its settings and make sure its server is reachable
  • Documents synced: At least one document must be indexed in your vector store

If configuration is missing, you'll see a prompt directing you to the appropriate settings page.