# Maya Insights > Maya Insights is a marketing analytics platform for e-commerce and marketing teams. It ingests advertising, web analytics, CRM, and e-commerce data into a governed warehouse and exposes it to AI assistants through a hosted Model Context Protocol (MCP) server, so users ask questions in natural language instead of building dashboards. The integration surface for AI agents is the Maya Insights MCP server. - Endpoint: `https://mcp.mayainsights.com/mcp` (Streamable HTTP) - Authentication: OAuth 2.0 authorization code with PKCE (S256) and dynamic client registration; public clients, no client secret. Protected-resource metadata is at `https://mcp.mayainsights.com/.well-known/oauth-protected-resource`, which names the authorization server advertising its authorize, token, registration, and JWKS endpoints. Access is per user and scoped to the organizations that user belongs to. Complete the OAuth flow; never ask a user to paste a credential. - Adding it: in any MCP-capable assistant, add the endpoint above as a custom connector. The same URL works everywhere. - A session lasts 30 days; access tokens are refreshed automatically within it. ## Tools The server exposes these tools. Prefer this order. - `snapshot` — the entry point for an analytics question. Pass the user's question in natural language. It selects the appropriate analysis, runs it, and returns a rendered result. That result is the deliverable; no follow-up routing call is needed. - `get_analysis_category` — lists the analysis categories the authenticated organization actually has. Use it for "what can I analyze?" questions, and whenever `snapshot` reports that no category matched. - `get_glossary` — describes the queryable data available within a category, in that organization's own terms. - `get_data` — runs a parameterized query against a category. Call `get_glossary` first. Use it for numbers a packaged analysis does not return. - `get_field_details` — explains what a single item returned by `get_glossary` means. - `submit_feedback` — files a bug report or feature request on the user's behalf. Show the user the draft and get explicit confirmation before calling it. - `list_feedback` — lists feedback the user has already submitted. Recommended call order: 1. Analytics question -> `snapshot`. 2. `snapshot` reports no matching category -> `get_analysis_category`, let the user choose, then continue with that category. 3. Numbers a packaged analysis does not cover -> `get_glossary` -> `get_data`. 4. Unclear what something means -> `get_field_details`. Every tool is read-only. What a given user can analyze depends on the data sources their organization has connected and on their subscription. Discover it at runtime with `get_analysis_category` rather than assuming a fixed list. If every tool errors at once, the connection has dropped — ask the user to reconnect Maya Insights from their account settings. ## Getting started for Shopify merchants 1. Install Maya Insights on your Shopify store. 2. Choose a plan and approve the charge in Shopify. 3. Your Maya Insights organization is provisioned automatically and your Shopify data starts syncing. 4. Copy your MCP URL from the post-purchase screen, or later from Settings in the Maya Insights app. 5. In Claude, open Settings -> Connectors -> Add custom connector, paste the URL, and sign in to Maya Insights to authorize the connection. 6. Open a new chat and ask a question about your store. Connecting more sources unlocks more analysis areas. ## Documentation - [Getting started](https://docs.mayainsights.com/docs/getting-started/): install to first answer. - [Connect Claude](https://docs.mayainsights.com/docs/connect-claude/): connector setup, how authentication works, revoking access. - [What you can analyze](https://docs.mayainsights.com/docs/what-you-can-analyze/): coverage and what unlocks each area. - [Example questions](https://docs.mayainsights.com/docs/example-questions/): questions that work well. - [MCP reference](https://docs.mayainsights.com/docs/mcp-reference/): endpoint, auth, and the tool contract. - [Troubleshooting](https://docs.mayainsights.com/docs/troubleshooting/). ## Optional - [Maya Insights website](https://www.mayainsights.com/) - [Maya Insights app](https://app.mayainsights.com/)