> ## 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.

# Pitch Audit write tools

> Create a Pitch Audit for a prospect, review and save its prompts, then confirm it through MCP. Requires an agency team on a plan with MCP writes.

These tools run the reviewable Pitch Audit lifecycle from an MCP client: create the audit, review the prompts, then confirm to start answer collection, which uses credits. Read [MCP write access](https://docs.qwairy.co/mcp/write/overview.md) first. Examples use synthetic data.

> [!NOTE]
>
> The Pitch Audit write tools require an agency team on an active Business, Agency Business or Enterprise plan. All three require the `write:pitch-audits` scope and the Manager role. Reading audits with [`list_pitch_audits` and `get_pitch_audit`](https://docs.qwairy.co/mcp/tools/pitch-audits) uses the separate `read:pitch-audits` scope.

| Tool | Confirmation | REST equivalent |
| - | - | - |
| `create_pitch_audit` | `cost_increase` | [`POST /pitch-audits`](https://docs.qwairy.co/developers/endpoints/pitch-audits/create) |
| `save_pitch_audit_prompts` | `always` | [`PATCH /pitch-audits/{auditId}/prompts`](https://docs.qwairy.co/developers/endpoints/pitch-audits/save-prompts) |
| `confirm_pitch_audit` | `always` | [`POST /pitch-audits/{auditId}/confirm`](https://docs.qwairy.co/developers/endpoints/pitch-audits/confirm) |

## create\_pitch\_audit

Create a Pitch Audit for a prospect. By default, prompts are prepared for review before any credit is used and the call needs no confirmation. When `autoStartResponses` is `true`, answer collection starts at once and uses credits, so the call needs a confirmation token.

### Parameters

| Parameter | Type | Required | Description |
| - | - | - | - |
| `teamId` | string | Yes | Agency team that owns the audit |
| `url` | string | Yes | Prospect website URL or domain, up to 2048 characters |
| `name` | string | No | Prospect display name, up to 120 characters |
| `countryCode` | string | No | Supported two-letter market code. The market controls model availability |
| `promptCount` | integer | No | Number of prompts to prepare, from 20 to 100. Default 30 |
| `customPrompts` | string\[] | No | 20 to 100 distinct prompt texts imported instead of generated prompts, each up to 1000 characters |
| `modelIds` | string\[] | No | Up to 20 models |
| `autoStartResponses` | boolean | No | Skips prompt review and starts answer collection, which uses credits and requires a confirmation token. Leave it unset to review first |
| `dryRun` | boolean | No | Preview without creating |
| `confirmationToken` | string | No | Token returned by the preview |

### Example

```text theme={null}
You: Prepare a Pitch Audit for prospect.example in the US with 20 prompts.

Assistant: calls create_pitch_audit with
  {"teamId": "team_agency_01", "url": "prospect.example", "countryCode": "US", "promptCount": 20}

Qwairy: creates the audit and queues prompt preparation. No credits are used
yet. The audit waits for review.
```

***

## save\_pitch\_audit\_prompts

Replace the prompts of a Pitch Audit that is awaiting review. Every call needs a confirmation token from a preview, because it replaces the saved prompts.

### Parameters

| Parameter | Type | Required | Description |
| - | - | - | - |
| `auditId` | string | Yes | Pitch Audit ID |
| `prompts` | object\[] | Yes | 20 to 100 prompts. Each has `id` (up to 50 characters), `text` (up to 1000), `category` (`TOFU`, `MOFU` or `BOFU`) and `topic` (up to 120) |
| `expectedVersion` | integer | Yes | The `version` of the audit you last read with `get_pitch_audit` |
| `dryRun` | boolean | No | Preview without saving |
| `confirmationToken` | string | No | Token returned by the preview |

The call replaces the whole prompt set. Send every prompt you want to keep.

### Example

```text theme={null}
You: Remove the two prompts about pricing and keep the rest.

Assistant: reads the audit (30 prompts), then calls save_pitch_audit_prompts
with the remaining 28 prompts and the audit version it read.
```

***

## confirm\_pitch\_audit

Confirm the prompts and start collecting answers. This uses the credits shown in the preview, so every execution needs a confirmation token.

### Parameters

| Parameter | Type | Required | Description |
| - | - | - | - |
| `auditId` | string | Yes | Pitch Audit ID |
| `prompts` | object\[] | No | 20 to 100 prompts to use, with the same fields as `save_pitch_audit_prompts` |
| `modelIds` | string\[] | No | Up to 20 models |
| `dryRun` | boolean | No | Preview the cost without starting |
| `confirmationToken` | string | No | Token returned by the preview |

### Example

```text theme={null}
You: Confirm the audit.

Assistant: calls confirm_pitch_audit with {"auditId": "audit_01"}

Qwairy: returns a preview with the credits the audit will use and a
confirmationToken. Nothing has started.

You: Confirm.

Assistant: repeats the call with the token. Answer collection starts, and the
assistant reads get_pitch_audit later to check the status.
```

Retrying a failed audit and exporting a report remain dashboard actions.

## Related

* [Run a Pitch Audit for a prospect with MCP](https://docs.qwairy.co/developers/guides/pitch-audit-prospect-mcp.md) for the full sequence.
* [Pitch Audit read tools](https://docs.qwairy.co/mcp/tools/pitch-audits.md).


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