The Playground is an interactive environment where developers can simulate and test conversations across voice, chat, and email before deploying their agents live. It lets you step into the shoes of an end user to see exactly how the AI agent responds across channels. It’s designed to catch bugs, fine-tune behavior, and validate flows under real-world conditions.

Yes. The Playground supports all major communication channels: you can switch between chat, voice, and email modes with a single click. This ensures your conversational design and tone are consistent across mediums and gives you full visibility into how your agent handles multimodal interactions.

The Playground includes a real-time Debug Panel. You can view the current and previous nodes, upcoming transitions, active agent logic, and the values of runtime variables. If a flow fails or gets stuck, you’ll see exactly which node caused it and why—helping you debug and resolve issues quickly.

Yes. During any test run, the Playground displays which AI agent or subflow is currently active, so you know exactly which logic is executing. If you’re chaining multiple agents or using fallback flows, this visibility helps ensure each one behaves as expected.

All runtime variables—like user name, issue type, session time, intent labels, API payloads—are visible in a structured panel. You can track how values are updated as the conversation progresses. This is critical for verifying if branching, conditions, and API calls are working as intended.

If an agent encounters an error or breaks during testing, the Playground highlights the exact step where it failed — whether it's a missing variable, API timeout, or intent mismatch. You can inspect the node, view the logs, and update the logic without redoing the whole flow.

Absolutely. The Playground operates in a safe, isolated test environment. None of the changes or test runs here affect live users. You can safely try new logic, simulate edge cases, and experiment freely before publishing.

Yes. In voice mode, the Playground simulates TTS-based calls with live transcription. In chat mode, you can type user inputs as if you're a real customer. This makes it easy to test how the AI agent responds to varied user behavior in real time.

Once your test cases pass, variables behave as expected, and all branches are verified across voice/chat/email—your AI agent is ready. The Playground acts as the final checkpoint before deployment, ensuring quality, reliability, and omnichannel consistency.