Skip to main content
Wiredesk is set up in the dashboard. As a developer, you usually have one of four jobs: connect your own systems to the workspace’s data, put the chat widget on a site and wire it into that site’s UI, let the agent call your own systems mid-conversation, or answer a security review. This tab covers all four.

​
What you can integrate with

​
Getting data out

  • Webhooks push each event to your server as it happens. See Webhooks.
  • The REST API reads the history whenever you ask, a page at a time. See REST API.
  • API call actions let the agent call your endpoint mid-conversation with the parameters you declared, so an order lookup can use your live data. See API calls.
  • CSV exports need nothing built: conversations, contacts, leads and extra fields each have one in the dashboard. See Exports.
Only /api/v1/… and /api/mcp are supported interfaces. The other /api/… routes the dashboard uses need a signed-in session and can change without notice. The widget’s events run in the visitor’s browser, not on your server. Use them for page-level analytics, and webhooks for a record of every conversation.

​
Docs for tools and AI assistants

The docs are published in plain-text forms too:
  • https://wiredesk.ai/docs/llms.txt is an index of every page in these docs, in the llms.txt format.
  • Add .md to the end of any docs page URL to get that page as Markdown, e.g. https://wiredesk.ai/docs/widget/install.md.
Point a coding assistant at these instead of the rendered HTML.

​
Developer pages

REST API

Keys, errors, pagination, and every endpoint.

Webhooks

Events on your server, signed and retried.

MCP server

The workspace as tools for an AI assistant.

Request logs

Every request your keys made in the last 30 days, with its status and error.

JavaScript API

Open, close and drive the widget from your own code.

Widget settings

Every wiredeskSettings field and data-* attribute, with defaults and precedence.

Events

What the widget emits, when, and with what payload.

Verified visitors

Sign a user id on your server so the agent knows who is chatting.

Security model

Keys, domains, identity, iframe isolation, rate limits and outbound calls.

Direct trunk

Route calls from your own SIP trunk to the agent.