









































Works out-of-box with your favorite AI agents

Seeq Analytics is an advanced analytics platform for process manufacturing time-series data. Makini's MCP server brings Seeq Analytics data and workflows to AI agents directly, with no custom integration to build.
Have any questions? We’re here to help You
Seeq is an analytics application, not a historian: it connects to historians such as AVEVA PI, Honeywell PHD, AspenTech IP.21, Emerson DeltaV and Ovation, GE Proficy, Wonderware and Ignition, and leaves the raw data where it is. Seeq Corporation, founded in Seattle in 2013, sells it to oil and gas, chemicals, pharma, mining, power and food producers. Engineers use Workbench for analysis, Organizer for reports, Data Lab for Python and Vantage for monitoring.
In Seeq a signal is a time ordered series of samples, such as a temperature tag with its unit and description. A capsule is a bounded period with a start and end time, and a condition is a set of capsules defined by one rule, for example every period when a pump ran above its vibration limit. Scalars are constants like thresholds, and assets group signals and conditions per equipment in an asset tree.
Yes, Seeq has a REST API, and the Seeq web application itself uses that API for everything it does; clients log in at the auth endpoint and receive a session token in the x-sq-auth header. Seeq generates SDKs from it for Python, Java and C#, with a limited MATLAB option, and ships the SPy library for Data Lab notebooks. A separate Connector SDK in C# and Java lets you add custom data sources.
Yes, Makini offers an MCP server for Seeq. You authorize your Seeq login through Makini Link, Makini issues an MCP endpoint, and you paste it into Claude, ChatGPT, Cursor or another MCP client. Call get_connection_details to see what the connection exposes; Makini's unified models are built around assets, work orders, meters and production data, so expect asset level and event level information rather than raw high frequency signals, and only where the connection supports it.
The Seeq AI Assistant lives inside Seeq: available to SaaS customers on R64 and later, it has General, Formula, Data Lab and Actions agents that explain Seeq Formula, write SPy code and create items in your worksheet, and you can plug in your own OpenAI or Azure OpenAI provider. An MCP agent through Makini sits outside Seeq, in Claude, ChatGPT or Cursor, and combines Seeq information with your CMMS or ERP through one endpoint.
An AI agent can use Seeq conditions as the trigger for maintenance work when Seeq and the CMMS are both connected through Makini: it reads capsules from a condition such as high bearing temperature, where exposed, matches the asset to the CMMS asset record, and drafts a work order where that connection and the user's rights allow writes. Keep a reliability engineer in the loop; a capsule is an event, not a fault.
Keep the Seeq connection read only unless there is a clear reason not to. Seeq items such as signals, conditions and formulas are built by engineers in Workbench, and a wrong edit changes every analysis that depends on them. Makini's tool list includes create and update tools, but a write succeeds only where the connection and the Seeq user's permissions allow it. Let the agent read, compare and summarize, and leave model changes to people.
An MCP connection to Seeq inherits the permissions of the Seeq user who authorizes it in Makini Link: Seeq applies access control to workbooks, folders and data source items, so the agent sees only what that user can open. Use a dedicated account with read access to the asset trees and workbooks in scope, and remember that Seeq itself only sees historians an administrator has connected. Every agent call through Makini is logged.