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

# Build with IQRush

> Use IQRush measurement data in an assistant, a reporting workflow, or your own application. These guides cover the MCP connector and the HTTP read interface used by configured integrations.

Use IQRush measurement data in an assistant, a reporting workflow, or your own application. These guides cover the MCP connector and the HTTP read interface used by configured integrations.

IQRush measures how brands and sources appear in AI answers. Its estimates, confidence intervals, and stability indicators help you decide whether a result is ready to support a decision. Preserve that context when you move the data into another system.

## Choose your starting point

| Your goal | Follow this path | First useful result |
| - | - | - |
| Connect ChatGPT or Claude | [Connect an assistant](/developers/guides/03-connect-assistants) → [first MCP calls](/developers/guides/02-mcp-quickstart) | A measurement from a selected, ready run. |
| Build an application | [Authorize your client](/developers/guides/04-authentication) → [check the HTTP connection](/developers/guides/06-http-quickstart) → [run the Python or JavaScript starter](/developers/guides/24-report-starter) | A JSON report from your own authorized integration. |
| Generate a report from an existing connection | [Choose a recipe](/developers/guides/16-report-recipes) → [check measurement quality](/developers/guides/10-stability) → [worked example](/developers/guides/17-worked-example) | An assessment with evidence, scope, and uncertainty. |

For a custom MCP client, use the [client guide](/developers/guides/05-custom-mcp-client).

## What the interfaces cover

The MCP tools read projects, runs, visibility measurements, supporting statements, source relationships, page information, and connected Google data. They also help organize an analysis. They do not provide project creation, measurement launch, Content Playground editing, or website publishing tools in this edition.

Agent Q is available inside IQRush. ChatGPT and Claude use the external MCP connection. Both approaches can help you explore measurement data; their available actions and conversation histories are separate. Do not assume a tool available to Agent Q is exposed to your external application.

Direct HTTP requests require a valid IQRush bearer credential for the environment you are using. The presence of a route in this reference does not create account access or a service-account entitlement. If you have only connected an assistant, continue through that assistant or implement the documented authorization flow for your own client.

## Know when to move to the next step

| Step | What you should have before continuing |
| - | - |
| Connect or authorize | The assistant can call IQRush, or your application has its own access token and matching HTTP base. |
| Select a project | A returned project ID for the intended workspace. |
| Select a measurement | A completed run ID and completion date, with data ready for the reads you need. |
| Read a result | The requested metric, source or brand, actual filters, value, interval, and quality fields. |
| Prepare a report | Scope and uncertainty carried into the output; limitations stated where quality or data is unavailable. |

Use one project, one run, and one metric for the first result. Once that path works, use the [MCP tool index](/developers/guides/23-tool-index) or [HTTP reference](/developers/guides/20-http-reference) to add the calls needed by your chosen recipe.


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