Atlas AI agent tools library
Atlas AI agents can interact with tools and functions, allowing them to perform complex tasks and interact with your applications. We provide built-in tools for Atlas AI agents, and you can also define custom tools to extend their capabilities.
Tools enable agents to move beyond simple text generation by providing access to data, computational capabilities, and external systems integration.
Query tools
The query tools use query features to retrieve and explore information from the Cognite knowledge graph.
Tool | Use to | Parameters |
---|---|---|
Query knowledge graph | Retrieve data instances from data models in Cognite Data Fusion (CDF). For high accuracy, configure the tool instructions with additional information on the selected data type. | Specify which data model, view, and credentials to use. |
Query time series data points | Retrieve data instances for data points. Use another tool to identify the relevant time series ID. See also: Time series in data modeling. | |
Find activities | Retrieve data instances for activities using the generic Query knowledge graph tool. The data model is Cognite core data model, and the view is set to Activity. | Specify which data model, view, and credentials to use. |
Find assets | Retrieve data instances for assets using the generic Query knowledge graph tool. The data model is Cognite core data model, and the view is set to Asset. | Specify which data model, view, and credentials to use. |
Find equipment | Retrieve data instances for equipment using the generic Query knowledge graph tool. The data model is Cognite core data model, and the view is set to Equipment. | Specify which data model, view, and credentials to use. |
Find files | Retrieve data instances for files using the generic Query knowledge graph tool. The data model is Cognite core data model, and the view is set to Files. See also: Files in data modeling. | Specify which data model, view, and credentials to use. |
Find maintenance orders | Retrieve data instances for maintenance orders using the generic Query knowledge graph tool. The data model is Cognite process industries data model, and the view is set to Maintenance order. | Specify which data model, view, and credentials to use. |
Find notifications | Retrieve data instances for notifications using the generic Query knowledge graph tool. The data model is Cognite process industries data model, and the view is set to Notifications. | Specify which data model, view, and credentials to use. |
Find operations | Retrieve data instances for operations using the generic Query knowledge graph tool. The data model is Cognite process industries data model, and the view is set to Operations. | Specify which data model, view, and credentials to use. |
Find time series | Retrieve data instances for time series using the generic Query knowledge graph tool. The data model is Cognite core data model, and the view is set to Time series. See also: Time series in data modeling. | Specify which data model, view, and credentials to use. |
Analysis tools
Tool | Use to | Parameters |
---|---|---|
Answer document questions | Get answers to questions from a file using semantic search. Use in combination with a query tool that retrieves files. See also: Files in data modeling and Optimizing data models for AI search. | |
Summarize documents | Summarize the information in specific files. Use in combination with a query tool that retrieves files. See also: Files in data modeling. | |
Analyze time series | Perform advanced calculations on time series data points. Use another tool to identify the relevant time series ID. See also: Time series in data modeling. |
Integration tools
The integration tools enable Atlas AI agents to interact with external systems and services.
beta
The integration tools are currently in beta testing and are subject to change.
Tool | Use to | Parameters |
---|---|---|
Call REST API | Call a Cognite API endpoint. | Specify the API endpoint to call and the method to use. |
Run Python code | Run custom Python code, and supply instructions for the code to run. | Specify the Python code to run or select an example script. |