Images in this article were generated by AI and may contain imperfections. Real application screenshots will be added soon.
This article may describe features that are still in development. If you need any of them prioritized, please contact us via the contact form (applies to Business subscription or higher).
How API and MCP Work
REST API: Token-authenticated endpoints provide access to sessions, analytics, recordings, and configuration. Use standard HTTP methods (GET, POST, PUT, DELETE) with JSON payloads. Rate limits vary by plan.
MCP Server: The Model Context Protocol allows AI assistants (Claude, ChatGPT, etc.) to query your analytics data directly. Instead of exporting data to a spreadsheet, you ask questions in natural language.
MCP connects via a simple configuration in your AI client. The server exposes tools for listing domains, fetching analytics, querying sessions, and analyzing patterns. All operations are read-only — AI cannot modify your configuration.
XLSX Export: For traditional reporting, the API supports XLSX export. Export analytics data, session lists, user journeys, and funnel events to spreadsheets compatible with Excel and Google Sheets.
Example MCP queries: "Show sessions with high bounce rate last week", "Which pages have the most rage clicks?", "Compare conversion rates between mobile and desktop".
![[img: horizontal illustration showing API endpoint documentation with code snippets, authentication flow, and response examples, developer documentation style]](/images/features/api1.webp)
Real-World Use Case
A product manager wants weekly UX insights but doesn't want to spend hours in dashboards. They configure the MCP server in Claude Desktop and ask: "Summarize user behavior issues from last week and suggest priorities."
Claude queries the API, analyzes rage click patterns, identifies pages with unusual drop-offs, and returns a prioritized list with supporting data. What took 2 hours of dashboard work now takes a 30-second conversation. For formal reports, they export data to XLSX and attach it to executive presentations.
![[img: horizontal illustration showing Claude Desktop chat interface with analytics query and response, MCP connection indicator visible, AI assistant conversation context]](/images/features/api2.webp)
![[img: horizontal illustration showing custom dashboard built with API data, multiple charts and session player embeds, data integration scenario]](/images/features/api3.webp)
Frequently Asked Questions
What's the difference between API and MCP?
The API is a traditional REST interface for programmatic access — useful for building custom integrations, dashboards, or automated workflows. MCP is a protocol that lets AI assistants use the API conversationally. Think of MCP as a natural-language wrapper around the API.
Which AI assistants support MCP?
Currently, Claude Desktop has native MCP support. Other assistants (ChatGPT, etc.) can be integrated via custom GPT configurations or similar mechanisms. The MCP server uses standard HTTP transport, so any system that can make authenticated requests can integrate.
Is the API rate limited?
Yes. Rate limits depend on your plan: Professional allows 100 requests/minute, Business allows 500 requests/minute, Enterprise has custom limits. Pagination is required for large result sets. Burst handling allows short-term overages.
Can AI modify my data through MCP?
No. The MCP server provides read-only access. AI can query analytics, list sessions, and fetch recording events, but cannot change settings, delete data, or modify configurations. This is a security design choice.
How do I set up the MCP server?
Generate an API token in account settings, then add the MCP configuration to your AI client (Claude Desktop settings file). The configuration specifies the server URL and your authentication token. Detailed setup instructions are in the dashboard.
What export formats are available?
The API supports JSON (default response format) and XLSX (spreadsheets). JSON is ideal for programmatic integrations, while XLSX is useful for reports, presentations, and analysis in Excel or Google Sheets. Choose the format via query parameter or Accept header.
