> ## Documentation Index
> Fetch the complete documentation index at: https://docs.condense.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Tool reference

> The 17 Condense MCP tools, what each one does, and which one changes your environment.

Condense MCP has 17 tools. Fifteen of them are in three permission groups, and you choose which to allow when you connect. The other two are internal and always on. You do not call the tools yourself: your AI tool picks them by reading their descriptions. This page lists them, so you know what you can ask for and which tools to leave switched on when you approve a connection or set up governance.

Your AI tool is the final word on what it can use. It shows the tools that your role, your approved permissions and your administrator allow, which can be fewer than the 17 below.

> [!NOTE]
>
> Pipeline and connector tools take a workspace name. Pod and job tools take a deployment or job name and its namespace. Name them in your request, so your AI tool has what the tool needs.

## Connectors and Logs

Six tools for pipelines, connector configurations, deploying connectors and connector logs.

* `Get-Current-Pipeline`: Gets the pipeline in a workspace when you give its name: which connectors it has, the connections between them and their resource usage.
* `Get-Deployed-Connector-Configs`: Gets the configuration of deployed connectors.
* `Get-Connector-Configs-To-Deploy`: Gets the configuration variables you need to deploy a connector.
* `Get-Custom-Connector-Configs-To-Deploy`: Gets the configuration variables you need to deploy a custom connector or application.
* `Deploy-connector`: Deploys a connector, prebuilt or custom. **This is the one tool that changes your environment.**
* `Get-Connector-Logs`: Gets the logs of a deployed connector. It needs the workspace name, the connector's configuration ID, connector ID or name, and how far back to read, in milliseconds.

## Metrics and Logs

Eight tools for utilization, workloads, metrics and alerts. They follow your Condense role like every other tool.

* `Get-CPU-Utilization`: Gets CPU utilization.
* `Get-Node-Utilization`: Gets node utilization.
* `Get-Pod-Status`: Gets the status of a pod when you give the deployment name and the namespace.
* `Get-Pod-Metrics`: Gets resource metrics when you give the deployment name and the namespace.
* `Read-Job`: Reads the details of a job when you give its name and namespace.
* `List-Metrics`: Lists the metrics available in Prometheus.
* `Execute-query`: Runs a PromQL query and returns the data.
* `Get-Latest-Alert`: Gets information about the latest Grafana alert that was triggered.

## Documentation

One tool for questions about Condense.

* `Query-docs`: Searches the Condense documentation to answer your questions about Condense.

## Internal tools

Two tools are used internally. You do not choose them when you approve a connection, and they cannot be disabled or paused. Your AI tool calls them on its own.

* `Get-all-environments`: Gets all the environments you have connected with your AI tool, across your connections.
* `Get-all-workspaces`: Gets all the workspaces you can access in an environment.

## The tool that changes your environment

`Deploy-connector` is the tool that deploys a connector. Three things to know before you ask your AI tool to use it:

* **Your role decides.** You need a Condense role that lets you deploy in that workspace. Approving a connection does not add to it.
* **An environment can block it.** If an administrator set the environment to Strict read-only, any action that changes state is blocked, whatever your role allows. See [Scope and permissions](https://docs.condense.io/condense/v3.0.0/mcp/scope-and-permissions.md).
* **You can leave it out.** Clear `Deploy-connector` on the approval page, or later on the AI Tools page, if an AI tool should only read.

Before you deploy, ask for the configuration variables first with `Get-Connector-Configs-To-Deploy` or `Get-Custom-Connector-Configs-To-Deploy`, and read the values your AI tool is about to send.

## Related

> [card] [Ask Condense MCP](https://docs.condense.io/condense/v3.0.0/mcp/ask-condense-mcp.md)
>
> Example requests for each tool, and what each one needs.

> [card] [Usage and limits](https://docs.condense.io/condense/v3.0.0/mcp/your-connected-ai-tools.md#usage-and-limits)
>
> The limits that apply to all tools, shared by your organization.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.