Playground

The Playground is an interactive experimentation area in the admin console. It lets space admins try all three core capabilities hands-on — without writing any code.

Chat

The Chat Playground provides an interactive conversation interface with:

• Streaming responses — See AI responses appear in real-time via SSE
• RAG toggle — Enable retrieval-augmented generation to ground responses in your documents
• Store selector — Choose which document store to search when RAG is enabled
• Temperature control — Adjust response creativity
• Conversation history — Right sidebar panel to switch between past conversations

Navigate to Playground > Chat in the sidebar.

Documents

The Documents Playground combines document management with semantic search:

• Drag-and-drop upload — Upload multiple files at once (PDF, DOCX, XLSX, CSV, Markdown, JSON, HTML, TXT)
• Store management — Select or create document stores from the header
• Document browser — Right sidebar shows uploaded documents with processing status
• Document preview — Click a document to see its extracted chunks
• Semantic search — Search across your documents with relevance scoring

Navigate to Playground > Documents in the sidebar.

Agents

The Agent Playground lets you run agent tasks and observe the reasoning loop:

• Goal input — Describe what the agent should accomplish
• Tool picker — Toggle which tools the agent can use (search, fetch, memory, custom skills)
• Live execution trace — Watch the agent’s think steps, tool calls, and responses in real-time
• Step truncation — Long outputs are collapsed with “Show more” toggles
• Studio integration — Open any session in Studio for deeper debugging
• Session history — Right sidebar to browse and reload past agent runs

Navigate to Playground > Agents in the sidebar.

From Playground to SDK

Everything you do in the Playground uses the same APIs the SDK uses. Once you’ve experimented and found the right configuration, you can replicate it in your backend with the Go SDK:

// Chat with RAG (mirrors Playground > Chat)
client.ChatCompletion(ctx, sdk.ChatCompletionRequest{
    Messages: []sdk.Message{{Role: "user", Content: "What is our refund policy?"}},
    UseRAG:   true,
    StoreID:  "store_abc123",
})

// Semantic search (mirrors Playground > Documents)
client.Search(ctx, sdk.SearchRequest{
    Query: "refund policy",
    TopK:  5,
})

// Run agent (mirrors Playground > Agents)
client.RunAgent(ctx, sessionID, "Summarize refund policies", nil)