Agent Playground

Schema Validation: Automatically checks tool definitions and input schemas.
Sandboxed Invocations: Calls are isolated from production environments.
Logging &amp; Telemetry: Response times, payloads, and errors are captured.
Multi-Interface Support: Test both chat-based and tool-only agents.
Streaming Preview: See live token-by-token responses (if supported).

The Agent Playground is a built-in testing environment where developers, administrators, and partners can experiment with agents in real time.

It provides a safe, isolated space to validate schemas, debug tools, and refine interactions before rolling agents into production.

Why Use the Playground?

Test Invocations – Run sample payloads against your agent’s /invoke endpoint.
Validate Metadata – Ensure your ai-tools.json conforms to the Agent Blueprints.
Check Health – Verify that your /health endpoint responds correctly.
Preview User Experience – See how agents behave when surfaced inside the Orbit Platform (chat, dashboards, or tools).
Debug Quickly – View structured logs, responses, and errors in one place.

Select or Load an Agent
- Choose a built-in, community, or locally running agent.
- Upload or link its ai-tools.json.
Send Test Payloads
- Use JSON-based request builder.
- Interact with tools, chat interfaces, or custom APIs.
Inspect Responses
- View raw JSON output.
- Preview how responses render in Orbit Flow components (e.g., chart, table, metric).
Iterate Quickly
- Modify schema, metadata, or prompts.
- Re-run tests until the agent behaves as expected.

Orbit Console → Navigate to Agent Services → Playground for development and testing.
Orbit Hosted → Playground available in your deployed platform for production testing.
CLI (Coming Soon)
API (Coming Soon)

Tip: The Playground is the fastest way to confirm your agent is blueprint-compliant and ready for Marketplace submission.