Dchat Widget Preview
Production-style embed sandbox
Launch-grade preview Cloud-hosted Dashboard Marketing site

Widget host

Preview the widget the way a paying customer will see it.

This host loads the production-style widget shell against the live API so you can judge brand fit, AI-first behavior, and handoff confidence before the script ever lands on a customer page.

AI-first replies Human takeover Starter prompts Paid-product polish
  • API base: https://api.dchat.com
  • Widget token: dchat-prod-2026
  • The launcher should appear in the bottom-right corner and respect the saved template, style, and launcher shape.
  • Use this page to test pricing questions, onboarding prompts, human handoff, and reopen after resolve.
Commercial brand fit The launcher, welcome line, and panel surface should feel like part of the site instead of a pasted-in support tool.
Operator confidence Every flow here should look deliberate enough that an agent could demo it live to a customer without apology.
10 launch templates Support, sales, onboarding, finance, healthcare, and more.
2 AI key models Bring your own GPT key or run on a Dchat-managed key.
1 shared inbox AI replies, human handoff, routing, and thread history stay together.

In context

Validate the launcher inside a realistic page shell

This host simulates the kind of marketing page or support page where the Dchat launcher will actually appear, so you can judge tone, spacing, and handoff flow together.

customer-site.example
AI-first support
Help visitors move from question to answer without losing the human option.

This mock shell exists to give the widget launcher a believable page context while you test starter prompts, welcome copy, and escalation timing.

Fast first response Let AI greet the visitor immediately and keep humans free for exceptions.
Clean escalation Ask for a person or trigger a routing rule to move the thread into the team queue.
Shared transcript The conversation stays intact across AI replies, handoff, resolve, and reopen states.

Embed

Drop this on the site

<script src="https://widget.dchat.com/loader.js"
                    data-api="https://api.dchat.com"
                    data-widget-token="dchat-prod-2026"></script>

Use the same token and API base shown above when you move from preview into a real embed context.

Suggested test path

Quick end-to-end pass

  1. Start a new thread from the widget launcher.
  2. Confirm the AI answer lands first.
  3. Ask for a human or trigger a routing keyword.
  4. Reply from the dashboard and watch the widget transcript refresh.