Skip to content

AI Chatbot

  • Managing Your Chatbot
  • Advanced
Guides
Choose an available topic from the sidebar. More documentation is coming soon.

Advanced

AI Chatbot Advanced Settings

MCP endpoints, user identity, CSV imports, CSP, and data handling for teams that need more control.

The AI chatbot advanced settings guides are for teams that have the basics working and need tighter control. They cover exposing a knowledge base to external AI agents with MCP, identifying logged-in users, preparing CSV data for a database tool, adding a Content Security Policy nonce, and reviewing how data is handled. Each guide is self-contained, so start with the one that matches your deployment question.

Expose a Knowledge Base via MCP

Every chatbot can act as a Model Context Protocol server. Use MCP endpoint to give tools like Claude, Cursor, or Windsurf a structured way to query trained content instead of scraping pages or the widget. The guide covers the endpoint, automatic WebMCP browser discovery, and the client setup.

Identify Users

Identify users links a conversation to a known account so answers and history can follow the person across sessions. The guide covers the identify call, the verification step, and the fields to pass for a reliable match.

Prepare Data with CSV to SQLite

CSV to SQLite turns a spreadsheet export into a queryable SQLite table for the SQLite action. Use it when the agent needs to look up rows from a catalog, schedule, or internal list that is easier to maintain as a CSV.

Content Security Policy

If your site sends a strict Content Security Policy, the widget script and its connections need to be allowed. Content Security Policy lists the directives and the nonce handling required to load the chatbot without weakening the rest of the policy.

Data Handling and Security

Data handling and security explains what the chatbot stores, how long conversations are retained, and where to find the GDPR and security references. Read it before a procurement or privacy review.