Product
Bring your AI assistant to your revenue data.
Connect a compatible AI client to DATALYR through MCP to investigate marketing performance, run supported reports and work with scoped tools. Use the workspace’s measurement context instead of pasting disconnected screenshots into chat.
Compare plans and usage
The assistant needs context as well as access.
An AI client can only answer reliably when it knows which workspace, date range and metric definition to use. DATALYR MCP gives compatible clients a structured way to access supported tools and data. The connection does not make every proposed action appropriate: review permissions, inspect results and preserve a clear boundary between analysis and applied changes.
In DATALYR
Marketing analytics MCP server
Connect through supported authentication
Use the documented hosted MCP setup and authorize the intended workspace. OAuth avoids turning a repository or chat transcript into a storage location for a secret key.
Discover the current tools
Let the client discover the available tool set at connection time. Supported capabilities can evolve, so a workflow should not depend on an old hard-coded tool list.
Review changes before application
MCP write tools preview changes by default. Read the proposed diff and destination before applying a supported write with the necessary scope.
The workflow
From setup to a useful answer.
Connect the client
Follow the current MCP setup instructions for your AI application. Confirm which workspace is accessible before asking a question involving business data.
Ask a bounded question
Specify the period and outcome you care about. Ask the assistant to show the measurement context and distinguish unavailable sources from measured zero results.
Retain useful work
Use supported reporting tools to preserve a useful query or analysis. For changes, review the dry-run result and confirm the intended effect before application.
Verify the result
Know what good looks like.
- The client can read the intended workspace context.
- The result includes an understandable period and metric basis.
- A write preview identifies the exact change and destination.
Go deeper
Use the guide for context, then follow the implementation reference for your setup.
A few practical questions.
Is MCP the same as the in-app assistant?
They are different access paths. MCP connects an external compatible client; Ask DATALYR is the conversational experience inside the application. Both should be used with clear measurement context.
Can an agent approve container scripts itself?
No. Container script proposals require owner or admin approval in the dashboard. A proposal is not an applied script, and API access does not remove that review boundary.
Explore related workflows.
- Marketing analytics API ↗
Read DATALYR marketing, attribution and revenue data through the HTTP API. Build internal workflows with explicit workspace access and metric definitions.
- AI marketing analyst ↗
Ask DATALYR questions about connected marketing and revenue data in plain English, then inspect the metrics and context behind the answer.
- Marketing reporting dashboards ↗
Build saved reports, dashboards and analyses in DATALYR so your team can review marketing and revenue performance with a consistent measurement basis.