# Run an insight from the command line

The plugin ships an `mixshift intelligence` command group so you can run a request without writing any code.

## The commands

- `mixshift intelligence catalog` lists the deployed requests: their id, and a one-line purpose for each.
- `mixshift intelligence run` runs one request against a merchant and period.
- `mixshift intelligence poll` and `mixshift intelligence get` check on or fetch an async run you started.
- `mixshift intelligence runs` lists the runs you have kicked off.

## An example run

A run needs the request id and its params: a merchant, and a period. For the ordered-sales bridge, comparing a month to the prior month:

```json
{
  "id": "INS-OPS-BRIDGE-01",
  "params": {
    "merchant": { "sellerId": "YOUR_SELLER_ID", "marketplaceId": "ATVPDKIKX0DER" },
    "period": { "month": "2026-07", "compare": "prior_month" }
  }
}
```

Swap `"compare": "prior_month"` for `"same_month_last_year"` to compare against a year ago instead, or pass explicit current and comparison date ranges if neither preset fits.

## The async flow

A large account can refuse an inline run on the retail-backed requests and ask you to retry with `"async": true`. Add that to the params, and the command hands back a run id instead of waiting. Use `mixshift intelligence poll` (or `get`, once it is finished) with that run id to check status and fetch the finished result. Each account can have two requests computing at a time; a third has to wait its turn.

## Where the result goes

A run writes its full result to a file and prints the headline straight to your terminal, so you get the quick answer immediately and the complete data to work with afterward.

## Related

- [What Intelligence answers](/knowledge-base/intelligence/getting-started/what-intelligence-answers)
- [Use Intelligence from your own code or agent](/knowledge-base/intelligence/how-to/use-intelligence-from-your-own-code-or-agent)
- [Limits and freshness](/knowledge-base/intelligence/reference/limits-and-freshness)