Besides looking at your dashboards, you can pull your attribution data into your own tools. The Analytics API is the way to do this with your own code, and the MCP server lets AI assistants such as Claude or ChatGPT query your data in a conversation.
In this article
1. Choose the right way to get your data
Option | What it does | Use it for |
Analytics API | You request KPIs and attribution data for a site, a preset and a date range whenever you need them | Data warehouses, BI tools, custom reports, scheduled exports |
MCP server | An AI assistant such as Claude or ChatGPT queries your Tracify data in a conversation | Ad-hoc questions, analyses and summaries without code |
CSV download | Export the recent events of a store under Settings › Activity | Checking single orders or events |
ℹ️ Note: Data from the last 24 hours is preliminary and can still change while it is processed. If you store data from the last 24 hours, request it again later to get the final numbers.
The Analytics API and the MCP server use the same API token. Sections 2 and 3 apply to both.
2. Create an API token
Open Settings › API Tokens in your Tracify dashboard.
Stay on the tab Analytics access and click Create analytics token. Tokens on the tab Server-side events are for sending events to Tracify, not for reading data.
Copy the token right away. Token values are only shown once.
The token gives read-only access to the analytics of every store the user account can access. Create a separate token per integration so you can revoke one without affecting the others.
3. Find your site ID and preset ID
Every request needs two IDs:
Site ID: identifies the store. You find it under Settings › General Settings in the box Store Information as Store ID.
Preset ID: identifies the attribution settings that are applied to the data, for example AI - All Touchpoints. You find it on every preset card under Settings › Presets. What a preset contains: How do presets work and which one should I use?
4. Query the Analytics API
The endpoints, parameters and responses are documented on tracify.dev/analytics-api. The page contains a Swagger UI where you can try every endpoint with your token:
Click Authorize and enter your API token.
Expand an endpoint to see its parameters and the response format.
Click Try it out, enter your site ID, preset ID and date range, and run the request.
The same documentation page lists the available levels of granularity, from channel totals down to single ads. For a daily sync, schedule the requests in your own system and request the previous day once it is complete. You might need the support of a developer.
5. Ask your data with an AI assistant (MCP server)
The Tracify MCP server connects your Tracify data to an AI assistant. Once connected, you ask questions in plain language, for example which campaigns had the best ROAS last week or how the new customer rate developed, and the assistant queries your Tracify data for you.
Works with Claude and ChatGPT, the setup takes a few minutes and needs your API token.
The assistant reads your analytics and customer journeys and can manage your influencers, cooperations and tracking links. It cannot change any other settings in Tracify.
You tell the assistant which store and which preset it should use, the same settings as in the dashboard.
Setup guide: How do I connect Tracify to Claude or ChatGPT (MCP)?
💡 Tip: For a quick check of single orders, the CSV download under Settings › Activity is enough. It lists the recent events of a store, including the order ID of purchases.
