> ## 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.

# Choose projects and measurement runs

> A project defines the market and questions you measure. A run is one measurement of that project. Keep the project and run IDs with every report so someone else can retrieve the same evidence.

A project defines the market and questions you measure. A run is one measurement of that project. Keep the project and run IDs with every report so someone else can retrieve the same evidence.

## Discover IDs

`list_projects` returns projects the caller can access. `get_project` returns project details and brand context. Use that context to check the brand terms and owned domains you intend to report.

`list_runs` returns runs newest first. Treat each `run_id` as an opaque string and pass it back unchanged. Current run-detail requests require a UUID. Do not convert the ID to a number or substitute a job ID.

## Check the run state

| Field | Meaning | Your next step |
| - | - | - |
| `status` | Execution state, such as `COMPLETED`, `FAILED`, `TERMINATED`, or `CANCELED`. | Select the completed run relevant to your question. |
| `completed_at` | When the measurement finished. | Use this date to label the measurement. |
| `data_ready` | Summary data is ready. | Still check the readiness of the data you request. |
| `cold_path_ready` | Grouped and filtered measurement data is ready. | If false, wait before requesting those views. If absent, readiness is unknown. |
| `data_readiness` from `get_run` | Status by available data family. | A missing family after completion can be unavailable or failed; completion does not guarantee every family exists. |
| `results_stability` from `get_run` | Not populated by this route. | Call `get_results_stability` separately. |

Do not treat a null count as zero. Some run-detail counts are unavailable even when useful measurements exist. Use a metric tool for the actual measurement rather than reconstructing it from run counters.

## Keep comparisons comparable

Use explicit run IDs for recurring reports. Record the selected topics, platforms, brand definitions, and filters. A changed project configuration can change the population being measured. A difference between two runs does not by itself establish that a content edit caused the difference.

`start_session` finds the most recent run, including one still in progress. It is useful for a conversation, but a scheduled reporting application should select and check its own run.

Next: [Choose filters and subjects](/developers/guides/08-filters-and-subjects).


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