Agent Playground
AVAILABLEThe Agent Playground is an interactive multi-turn testing environment for debugging AI agents, inspecting reasoning chains, and validating tool execution.
- Portal Location:
Projects > [Your Project] > AI & Capabilities > Agent Playground(/projects/[slug]/playground)
Key Features #
- Multi-Turn Chat History: Test complex conversational state across multiple turns.
- Real-Time Tool Execution Inspection: View exact JSON parameters passed to registered Tools and external MCP servers.
- Step-by-Step Reasoner Logs: Inspect DeepSeek-R1 / OpenAI reasoning traces and intermediate thought chains.
- Instant Parameter Tweaking: Dynamically adjust temperature, top_p, and active tools without restarting sessions.
Rendering diagram...
sequenceDiagram
autonumber
actor Dev as Developer in Playground
participant Agent as Naagmani Agent Runtime
participant Tool as Registered MCP Tool (SQL Database)
Dev->>Agent: "Find the top 3 users by token usage this month"
Agent->>Agent: Generate Plan & Tool Call (query_sql)
Agent->>Tool: Execute query_sql({"query": "SELECT * FROM..."})
Tool-->>Agent: Returns Rows JSON
Agent->>Dev: Stream Formatted Markdown Response + Inspectable Tool Trace
How to Test an Agent #
- Select your target agent from the Agent Selector dropdown.
- Enter your message in the chat input.
- Watch the assistant response stream in real time.
- Expand the Tool Trace Drawer to inspect input arguments, returned payloads, and execution latency.