Skip to main content
The Playground is the safest place to test how your bot behaves before you put it in front of customers. It uses the live WebChat pipeline, not a fake demo responder, so you see exactly what end users will see.

​
Before you can use it

The Playground only shows working sessions for bots that already have a WebChat channel with a public key. If no WebChat channel exists yet, the page shows a shortcut to Channels -> WebChat so you can create one first.
You need at least one WebChat channel connected to a bot before Playground is available. Create the channel first, then return to Playground to start testing.

​
How testing works

When you open Playground, BoundBot:
  1. lets you choose from bots that already have WebChat channels
  2. loads that channel’s widget configuration
  3. starts a chat session for the current browser
  4. loads recent message history for that session
  5. streams replies the same way the embedded widget does
This makes Playground useful for checking real bot behavior before you embed or publish a channel.

​
Interact with your bot in real time

Playground is the right place to validate how your bot responds before going live. Test system prompts and tone Change the bot’s system prompt in Bots (open in app), then switch back to Playground and send a few messages. You can confirm that instructions about tone, language, and boundaries are reflected in the replies. Iterate on the prompt until the bot sounds right for your brand. Test knowledge retrieval Ask questions that should be answered from your Knowledge base—FAQs, file content, website snippets, or product data. Playground uses the same retrieval pipeline as production, so you can verify that the bot finds the right sources and cites them appropriately. If an answer is wrong or missing, update the knowledge source and test again in Playground.
Run through a short list of real customer questions in Playground after every knowledge or prompt change. Catching issues here saves time and keeps customer-facing answers accurate.

​
What to test here

  • welcome messages and tone
  • FAQ, file, website, and product answers
  • action and workflow triggers
  • retry behavior when a response fails
  • stop behavior while a reply is streaming

​
Troubleshooting and debugging

Use a fresh session when needed Playground ties a session to your browser. If you want to simulate a brand-new visitor (e.g. to test the welcome message or first-run flow), open an incognito or private window, or use a different browser profile. That way you avoid reusing the same stored chat session and see the same experience as a first-time user. Check channel status If Playground does not load or the bot does not reply, confirm that the WebChat channel is active and assigned to the correct bot. In Channels -> WebChat, check that the channel shows a healthy status and that the bot you selected in Playground is the one connected to that channel. Verify knowledge and bot settings If answers are missing or off, double-check that the bot has access to the right knowledge sources and that those sources are published. Also confirm the bot’s system prompt and any workflow or action triggers that might affect replies.
Playground uses your live bot and real credits. Test with short, focused conversations so you can iterate quickly without burning through your quota.

​
Practical testing tips

Use Playground before and after every major bot or knowledge change. A quick run-through here helps you ship with confidence.
Bookmark the Playground URL and run a few test questions after each deployment or content update. Consistent use keeps your bot reliable for customers.

WebChat

Create or edit the WebChat channel that powers Playground.

Bots

Update the bot that replies in Playground.

Inbox

Confirm the same conversations appear in the shared inbox.

Embed WebChat

Move from internal testing to a real website embed.