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

# Run write tool

> Start a run through MCP to collect answers now for selected prompts and models. The tool reserves credits and always needs a preview and confirmation token.

A run collects answers for the selected prompts and models immediately instead of waiting for the monitoring schedule. Read [MCP write access](https://docs.qwairy.co/mcp/write/overview.md) first for plans, scopes, roles and confirmation tokens. Examples use synthetic data.

| Tool | Scope | Minimum role | Confirmation | REST equivalent |
| - | - | - | - | - |
| `start_run` | `write:monitoring` | Manager | `always` | [`POST /brands/{brandId}/runs`](https://docs.qwairy.co/developers/endpoints/runs/manage) |

## start\_run

Collect answers now for the selected prompts and models. The tool reserves the credits shown in the preview, so it needs a confirmation token on every execution.

### Parameters

| Parameter | Type | Required | Description |
| - | - | - | - |
| `brandId` | string | Yes | ID of a `LIVE` workspace |
| `promptIds` | string\[] | No | Prompts to run. Defaults to every active prompt in the workspace |
| `modelIds` | string\[] | No | Models to run. Defaults to the workspace monitoring models |
| `dryRun` | boolean | No | Preview the targets and cost without reserving credits |
| `confirmationToken` | string | No | Token returned by the preview |

### Before you start

1. Call [`estimate_run`](https://docs.qwairy.co/mcp/tools/runs#estimate_run) to see the usable prompts, models and the estimated credits.
2. Call [`get_credits`](https://docs.qwairy.co/mcp/tools/runs#get_credits) if you need to check that the balance covers the cost.
3. Call `start_run` without a token to receive the preview and the token, then repeat the call with the token.

### After the run starts

Collection is asynchronous. Follow progress with [`get_run` and `list_runs`](https://docs.qwairy.co/mcp/tools/runs). There is no MCP tool to cancel or retry a run. To be notified by your own systems when a run finishes, subscribe to the [`run.completed` event](https://docs.qwairy.co/developers/integration-webhooks/events#run-completed) with the REST API.

> [!NOTE]
>
> The preview shows the credits that would be reserved. Moving credits between teams and buying more remain dashboard actions.

### Example

```text theme={null}
You: Run the five CRM prompts for Acme Analytics on ChatGPT and Gemini now.

Assistant: calls estimate_run, which reports 10 answers and the credits needed,
then calls start_run with
  {"brandId": "brand_01", "promptIds": ["prompt_01", "prompt_02", "prompt_03", "prompt_04", "prompt_05"],
   "modelIds": ["openai-gpt", "google-gemini"]}

Qwairy: returns a preview of the targets and the credits to reserve, and a
confirmationToken. Nothing has started.

You: Confirm.

Assistant: calls start_run again with the same arguments and the token, then
reads get_run to follow progress.
```

## Related

* [Run and credit tools](https://docs.qwairy.co/mcp/tools/runs.md) for estimates, status and the credit balance.
* [Onboard a client with the API](https://docs.qwairy.co/developers/guides/onboard-client-api.md) for a REST workflow that ends with a first run.


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