Using the Deploy MCP tools

You can access the following Deploy MCP tools from your MCP client:

Tools	Description	Input	Additional Info
list_applications	Lists all applications.	None	You must use the list_applications tool first to get the application IDs before using any other deployment functions.
list_application_processes	Lists all processes defined in an application.	`applicationId`	None.
list_application_environments	Lists the environments for a specific application.	`applicationId`	You must use the list_application_environments tool after the list_applications tool to get the environment IDs before deploying.
list_application_snapshots	Lists all snapshots defined in an application.	`applicationId`	None.
get_deployment_status	Returns the application deployment status.	`deploymentId`	None.
deploy_snapshot_to_environment	Deploys an application snapshot to an environment.	`applicationId`, `snapshotId`, `processId`, `environmentId`	None.
schedule_deployment	Schedule a deployment for future execution.	`deploymentId`	None.
get_deployment_failed_status	Retrieves information about the failed steps in a deployment.	`deploymentId`	The step produces a step output token ID or log ID which is used by the get_deployment_step_output and search_step_output_regex tools.
get_deployment_step_output	Retrieves information on a step in a deployment.	`stepOutputTokenOpaque`, `firstLineIndex`	The input `firstLineIndex` value as `-50` reads the last 50 lines of log. You must call the get_deployment_failed_status tool first to get the required step output token ID or log ID.
search_step_output_regex	Searches logs using regular expressions.	`stepOutputTokenOpaque`, `pattern`, `contextLineCount`	Reads the specific log segment, and lists the error messages that match the pattern. You must call the get_deployment_failed_status tool first to get the required step output token ID or log ID. The tool utilizes regex search (RE2 engine) with context lines to diagnose issues like checksum failures or module download errors.
search_step_output_hangs	Identifies long intervals (hangs) between output lines in step execution logs by analyzing timestamp information in the log format. Currently, this tool only checks for hangs within failed steps.		It returns matching hangs with their duration and surrounding context, helping diagnose deployment steps that stall or take unexpectedly long.

Tip: You can retrieve a complete list of all available tools directly from your MCP client. For specific instructions on how to use your MCP client to view this list, refer to the independent software vendor documentation.

Note: The additional information provided in the table is for general reference only. For programmatic integration, you must always rely on the self-describing schemas provided by the MCP server.

The core functionality of the Deploy MCP tools is accessed through the chat prompts within the MCP client interface. By using these prompts, you can perform complex deployment tasks by using natural language. Examples of the chat prompts that you can use are as follows:

List my applications.
List processes for the application helloApplication.
Deploy helloApplication into dev_environment1.
Why did deployment deployment_id fail?

Note: When crafting your prompts, you can reference objects by their ID or name.

The AI agent interprets your chat prompts to intelligently select and run the appropriate MCP tool to fulfill the request.

Important: The results are highly dependent on the capabilities of the LLM that your MCP client uses, and the results are non-deterministic.