---
title: "Formats and prompts"
description: "Article formats (the SEO blueprints the AI follows) and each pipeline stage's prompt.  Read the defaults, override one per Site, or reset it."
canonical: "https://writavo.com/docs/api/formats-and-prompts"
last-updated: "2026-09-30"
---

# Formats and prompts

Article formats (the SEO blueprints the AI follows) and each pipeline stage's prompt.

Read the defaults, override one per Site, or reset it.

Base URL: `https://api.writavo.com/v1`

### POST /formats

Define an article format

- **Operation id**: `createFormat`
- **Scope**: `prompts:write`
- **Permission**: `prompts.manage`
- **Rate limit class**: write
- **Plan feature**: `none`
- **Spends credits**: no
- **Publishable key may call it**: no

Define a format for this Site: a name, a key, and a blueprint of the sections the generator
writes and the structured blocks it must include. A key equal to a platform default's overrides
that default for this Site (only the override is listed afterwards); a key another Site format
already uses is `409 CONFLICT`.

`blueprint.requires` names the structured blocks the generator must fill: `howto_steps`,
`key_takeaways`, `comparison`, `faqs`. `blueprint.sections` is the outline, in order.

**Parameters**

| Name | In | Type | Required | Description |
|---|---|---|---|---|
| `Idempotency-Key` | header | string | yes | A unique key per logical operation, 1 to 255 printable ASCII characters (a UUID works), scoped to the Site and endpoint and kept for 24 hours. - Same key and body: the original response is replayed, so retrying a timeout is safe. - Same key, different body: `409 IDEMPOTENCY_KEY_CONFLICT`. - Same key while the first is running: `409 IDEMPOTENCY_KEY_IN_FLIGHT`; wait and retry. |

**Request body** (required)

| Field | Type | Required | Description |
|---|---|---|---|
| `name` | string | yes |  |
| `key` | string | no | Derived from `name` when omitted. A platform default's key overrides that default for this Site. |
| `blueprint` | Blueprint | yes | What a format tells the generator: the outline to follow and the structured blocks to fill. |
| `is_active` | boolean | no |  |
| `description` | string | null | no |  |
| `prompt_addendum` | string | null | no | Writing instructions for this format. May use the AI writing profile placeholders. |

**Responses**

| Status | Data | Description |
|---|---|---|
| 201 | FormatDetail | The new format. |
| 400 | - | `INVALID_REQUEST`. Bad JSON, an unusable query parameter value, or a missing required header. |
| 401 | - | `INVALID_API_KEY`, `API_KEY_REVOKED` or `API_KEY_EXPIRED`: no key, or an unusable one. It never distinguishes "no such key" from "wrong key". |
| 403 | - | `INSUFFICIENT_SCOPE`. The key lacks this operation's scope, or its creator's permissions no longer cover it. Another Site's object is always 404, never this. A 403 may also be: - `AGENT_ACCESS_DISABLED` - AI agent access is turned off. - `APPROVAL_DENIED` - a person denied this action. - `FORBIDDEN` (organisation, team, settings) - a change no scope allows, such as an AI agent changing the access of the person it acts for, anything about ownership, or a setting only a person may change. The message says what to do instead. |
| 409 | - | The request collides with the current state. Read `code`: - `SLUG_CONFLICT` - the slug is taken. - `IDEMPOTENCY_KEY_CONFLICT` / `IDEMPOTENCY_KEY_IN_FLIGHT` - key reused with another body, or still running. - `CONFLICT` - the object changed under you. Re-read and retry. - `APPROVAL_INVALID` - the `Writavo-Approval` is expired, used, unknown, for another request, or its effect changed. Retry without it for a new approval. - `PREREQUISITE_MISSING` / `FEATURE_UNAVAILABLE` (settings, delivery, SEO) - something is not set up (the message names what sets it up), or Writavo has it switched off (the message names the free alternative). Nothing was done or charged. |
| 422 | - | `VALIDATION_FAILED`. The values are not acceptable; `error.fields` maps each field to a message. Common causes: publishing without a title, slug or content; scheduling in the past; sending `status` on a create or update. |
| 429 | - | `RATE_LIMIT_EXCEEDED`. Back off and honour `Retry-After`. Limits are per key, per minute, by endpoint class. The two device sign-in endpoints take no key, so they are limited per IP address. \| Class \| Endpoints \| Limit \| \|---\|---\|---\| \| read \| every GET \| 600 per minute \| \| write \| POST, PATCH and DELETE on content, taxonomy, keys, webhooks, the content plan, formats and prompts, settings, delivery, SEO curation, the team and billing \| 120 per minute \| \| upload \| `POST /media/upload-url` and `POST /seo/backlinks/import` \| 60 per minute \| \| pipeline \| `POST /pipeline/runs`, `POST /pipeline/articles/{id}/requeue`, `POST /delivery/cms/push` and the seven paid SEO scans \| 10 per minute \| \| auth \| `POST /auth/device` and `POST /auth/device/token`, per IP address \| 60 per minute \| `POST /comments` adds three limits, named in `error.limit` on its 429: `reader` (5 per reader in 10 minutes, by `reader_ip`), `site` (300 per Site per hour) and `queue` (2,000 awaiting review). Every response carries `RateLimit-Limit`, `RateLimit-Remaining` and `RateLimit-Reset`. |
| 500 | - | `INTERNAL_ERROR`. Our fault. Quote the `request_id` from the body if you contact support. Safe to retry; send the same `Idempotency-Key` so a request that did succeed is not repeated. |
| 503 | - | `MAINTENANCE`. Writes are paused, platform wide or for your Site. Reads usually work and your published blog stays up. Retry after `Retry-After`. |

```bash
curl -X POST https://api.writavo.com/v1/formats \
  -H "Authorization: Bearer wv_sk_EXAMPLE0000000000000000000000000000" \
  -H "Idempotency-Key: $(uuidgen)" \
  -H "Content-Type: application/json" \
  -d '{
  "name": "<name>",
  "blueprint": "<blueprint>"
}'
```

### GET /pipeline/prompts

Read the AI prompts

- **Operation id**: `listPrompts`
- **Scope**: `pipeline:read`
- **Permission**: `none`
- **Rate limit class**: read
- **Plan feature**: `none`
- **Spends credits**: no
- **Publishable key may call it**: no

The prompt used at each of the five pipeline stages that read one (`plan`, `generate`,
`improve`, `classify`, `extract`): the platform default, this Site's override if it has one,
which of the two is in effect, and the `{{placeholders}}` the default uses.

Needs `pipeline:read`, so a read-only connection can see which prompts the Site uses and
whether it overrides any. The template TEXT (the default's and the override's) is included
only for a caller that may manage prompts (a key carrying `prompts:write`, or a signed-in
member with that permission); for anyone else both are `null` and `templates_included` is
`false`.

**Responses**

| Status | Data | Description |
|---|---|---|
| 200 | object | Every stage's prompt. |
| 401 | - | `INVALID_API_KEY`, `API_KEY_REVOKED` or `API_KEY_EXPIRED`: no key, or an unusable one. It never distinguishes "no such key" from "wrong key". |
| 403 | - | `INSUFFICIENT_SCOPE`. The key lacks this operation's scope, or its creator's permissions no longer cover it. Another Site's object is always 404, never this. A 403 may also be: - `AGENT_ACCESS_DISABLED` - AI agent access is turned off. - `APPROVAL_DENIED` - a person denied this action. - `FORBIDDEN` (organisation, team, settings) - a change no scope allows, such as an AI agent changing the access of the person it acts for, anything about ownership, or a setting only a person may change. The message says what to do instead. |
| 429 | - | `RATE_LIMIT_EXCEEDED`. Back off and honour `Retry-After`. Limits are per key, per minute, by endpoint class. The two device sign-in endpoints take no key, so they are limited per IP address. \| Class \| Endpoints \| Limit \| \|---\|---\|---\| \| read \| every GET \| 600 per minute \| \| write \| POST, PATCH and DELETE on content, taxonomy, keys, webhooks, the content plan, formats and prompts, settings, delivery, SEO curation, the team and billing \| 120 per minute \| \| upload \| `POST /media/upload-url` and `POST /seo/backlinks/import` \| 60 per minute \| \| pipeline \| `POST /pipeline/runs`, `POST /pipeline/articles/{id}/requeue`, `POST /delivery/cms/push` and the seven paid SEO scans \| 10 per minute \| \| auth \| `POST /auth/device` and `POST /auth/device/token`, per IP address \| 60 per minute \| `POST /comments` adds three limits, named in `error.limit` on its 429: `reader` (5 per reader in 10 minutes, by `reader_ip`), `site` (300 per Site per hour) and `queue` (2,000 awaiting review). Every response carries `RateLimit-Limit`, `RateLimit-Remaining` and `RateLimit-Reset`. |
| 500 | - | `INTERNAL_ERROR`. Our fault. Quote the `request_id` from the body if you contact support. Safe to retry; send the same `Idempotency-Key` so a request that did succeed is not repeated. |
| 503 | - | `MAINTENANCE`. Writes are paused, platform wide or for your Site. Reads usually work and your published blog stays up. Retry after `Retry-After`. |

```bash
curl https://api.writavo.com/v1/pipeline/prompts \
  -H "Authorization: Bearer wv_sk_EXAMPLE0000000000000000000000000000"
```

### PATCH /pipeline/prompts/{stage}

Override a stage's prompt

- **Operation id**: `setPromptOverride`
- **Scope**: `prompts:write`
- **Permission**: `prompts.manage`
- **Rate limit class**: write
- **Plan feature**: `none`
- **Spends credits**: no
- **Publishable key may call it**: no

Replace the prompt this Site uses at one stage. The override takes effect on the next article
written at that stage; articles already written are unchanged. Each save is a new version.

Keep the default's `{{placeholders}}`: they are how the stage passes in the topic, the knowledge
profile and the research. A placeholder the override leaves out is listed in
`missing_placeholders` as a warning; the save still happens, and the stage then works without
that input. `DELETE /pipeline/prompts/{stage}` goes back to the default.

**Parameters**

| Name | In | Type | Required | Description |
|---|---|---|---|---|
| `stage` | path | PromptStage | yes | The pipeline stage whose prompt this is. |

**Request body** (required)

| Field | Type | Required | Description |
|---|---|---|---|
| `template` | string | yes |  |

**Responses**

| Status | Data | Description |
|---|---|---|
| 200 | PromptSaved | The stage's prompt after the change. |
| 400 | - | `INVALID_REQUEST`. Bad JSON, an unusable query parameter value, or a missing required header. |
| 401 | - | `INVALID_API_KEY`, `API_KEY_REVOKED` or `API_KEY_EXPIRED`: no key, or an unusable one. It never distinguishes "no such key" from "wrong key". |
| 403 | - | `INSUFFICIENT_SCOPE`. The key lacks this operation's scope, or its creator's permissions no longer cover it. Another Site's object is always 404, never this. A 403 may also be: - `AGENT_ACCESS_DISABLED` - AI agent access is turned off. - `APPROVAL_DENIED` - a person denied this action. - `FORBIDDEN` (organisation, team, settings) - a change no scope allows, such as an AI agent changing the access of the person it acts for, anything about ownership, or a setting only a person may change. The message says what to do instead. |
| 422 | - | `VALIDATION_FAILED`. The values are not acceptable; `error.fields` maps each field to a message. Common causes: publishing without a title, slug or content; scheduling in the past; sending `status` on a create or update. |
| 429 | - | `RATE_LIMIT_EXCEEDED`. Back off and honour `Retry-After`. Limits are per key, per minute, by endpoint class. The two device sign-in endpoints take no key, so they are limited per IP address. \| Class \| Endpoints \| Limit \| \|---\|---\|---\| \| read \| every GET \| 600 per minute \| \| write \| POST, PATCH and DELETE on content, taxonomy, keys, webhooks, the content plan, formats and prompts, settings, delivery, SEO curation, the team and billing \| 120 per minute \| \| upload \| `POST /media/upload-url` and `POST /seo/backlinks/import` \| 60 per minute \| \| pipeline \| `POST /pipeline/runs`, `POST /pipeline/articles/{id}/requeue`, `POST /delivery/cms/push` and the seven paid SEO scans \| 10 per minute \| \| auth \| `POST /auth/device` and `POST /auth/device/token`, per IP address \| 60 per minute \| `POST /comments` adds three limits, named in `error.limit` on its 429: `reader` (5 per reader in 10 minutes, by `reader_ip`), `site` (300 per Site per hour) and `queue` (2,000 awaiting review). Every response carries `RateLimit-Limit`, `RateLimit-Remaining` and `RateLimit-Reset`. |
| 500 | - | `INTERNAL_ERROR`. Our fault. Quote the `request_id` from the body if you contact support. Safe to retry; send the same `Idempotency-Key` so a request that did succeed is not repeated. |
| 503 | - | `MAINTENANCE`. Writes are paused, platform wide or for your Site. Reads usually work and your published blog stays up. Retry after `Retry-After`. |

```bash
curl -X PATCH https://api.writavo.com/v1/pipeline/prompts/3f1b0c7a-0000-4000-8000-000000000001 \
  -H "Authorization: Bearer wv_sk_EXAMPLE0000000000000000000000000000" \
  -H "Content-Type: application/json" \
  -d '{
  "template": "<template>"
}'
```

### DELETE /pipeline/prompts/{stage}

Reset a stage's prompt to the default

- **Operation id**: `resetPrompt`
- **Scope**: `prompts:write`
- **Permission**: `prompts.manage`
- **Rate limit class**: write
- **Plan feature**: `none`
- **Spends credits**: no
- **Publishable key may call it**: no

Remove this Site's override for one stage, so the platform default is used again from the next
article. `404` when the stage has no override.

**Parameters**

| Name | In | Type | Required | Description |
|---|---|---|---|---|
| `stage` | path | PromptStage | yes | The pipeline stage whose prompt this is. |

**Responses**

| Status | Data | Description |
|---|---|---|
| 204 | - | Reset. No body. |
| 401 | - | `INVALID_API_KEY`, `API_KEY_REVOKED` or `API_KEY_EXPIRED`: no key, or an unusable one. It never distinguishes "no such key" from "wrong key". |
| 403 | - | `INSUFFICIENT_SCOPE`. The key lacks this operation's scope, or its creator's permissions no longer cover it. Another Site's object is always 404, never this. A 403 may also be: - `AGENT_ACCESS_DISABLED` - AI agent access is turned off. - `APPROVAL_DENIED` - a person denied this action. - `FORBIDDEN` (organisation, team, settings) - a change no scope allows, such as an AI agent changing the access of the person it acts for, anything about ownership, or a setting only a person may change. The message says what to do instead. |
| 404 | - | `NOT_FOUND`. No such object, or it belongs to another Site. **The two are deliberately indistinguishable and never return 403.** If the id is right, check you are using the right Site's key. |
| 429 | - | `RATE_LIMIT_EXCEEDED`. Back off and honour `Retry-After`. Limits are per key, per minute, by endpoint class. The two device sign-in endpoints take no key, so they are limited per IP address. \| Class \| Endpoints \| Limit \| \|---\|---\|---\| \| read \| every GET \| 600 per minute \| \| write \| POST, PATCH and DELETE on content, taxonomy, keys, webhooks, the content plan, formats and prompts, settings, delivery, SEO curation, the team and billing \| 120 per minute \| \| upload \| `POST /media/upload-url` and `POST /seo/backlinks/import` \| 60 per minute \| \| pipeline \| `POST /pipeline/runs`, `POST /pipeline/articles/{id}/requeue`, `POST /delivery/cms/push` and the seven paid SEO scans \| 10 per minute \| \| auth \| `POST /auth/device` and `POST /auth/device/token`, per IP address \| 60 per minute \| `POST /comments` adds three limits, named in `error.limit` on its 429: `reader` (5 per reader in 10 minutes, by `reader_ip`), `site` (300 per Site per hour) and `queue` (2,000 awaiting review). Every response carries `RateLimit-Limit`, `RateLimit-Remaining` and `RateLimit-Reset`. |
| 500 | - | `INTERNAL_ERROR`. Our fault. Quote the `request_id` from the body if you contact support. Safe to retry; send the same `Idempotency-Key` so a request that did succeed is not repeated. |
| 503 | - | `MAINTENANCE`. Writes are paused, platform wide or for your Site. Reads usually work and your published blog stays up. Retry after `Retry-After`. |

```bash
curl -X DELETE https://api.writavo.com/v1/pipeline/prompts/3f1b0c7a-0000-4000-8000-000000000001 \
  -H "Authorization: Bearer wv_sk_EXAMPLE0000000000000000000000000000"
```

### GET /formats/{id}

Read one article format

- **Operation id**: `getFormat`
- **Scope**: `meta:read`
- **Permission**: `none`
- **Rate limit class**: read
- **Plan feature**: `none`
- **Spends credits**: no
- **Publishable key may call it**: no

One format with its blueprint: the sections the generator writes and the structured blocks it
must include. Platform defaults and Site formats are both readable.

**Parameters**

| Name | In | Type | Required | Description |
|---|---|---|---|---|
| `id` | path | uuid | yes |  |

**Responses**

| Status | Data | Description |
|---|---|---|
| 200 | FormatDetail | The format. |
| 401 | - | `INVALID_API_KEY`, `API_KEY_REVOKED` or `API_KEY_EXPIRED`: no key, or an unusable one. It never distinguishes "no such key" from "wrong key". |
| 403 | - | `INSUFFICIENT_SCOPE`. The key lacks this operation's scope, or its creator's permissions no longer cover it. Another Site's object is always 404, never this. A 403 may also be: - `AGENT_ACCESS_DISABLED` - AI agent access is turned off. - `APPROVAL_DENIED` - a person denied this action. - `FORBIDDEN` (organisation, team, settings) - a change no scope allows, such as an AI agent changing the access of the person it acts for, anything about ownership, or a setting only a person may change. The message says what to do instead. |
| 404 | - | `NOT_FOUND`. No such object, or it belongs to another Site. **The two are deliberately indistinguishable and never return 403.** If the id is right, check you are using the right Site's key. |
| 429 | - | `RATE_LIMIT_EXCEEDED`. Back off and honour `Retry-After`. Limits are per key, per minute, by endpoint class. The two device sign-in endpoints take no key, so they are limited per IP address. \| Class \| Endpoints \| Limit \| \|---\|---\|---\| \| read \| every GET \| 600 per minute \| \| write \| POST, PATCH and DELETE on content, taxonomy, keys, webhooks, the content plan, formats and prompts, settings, delivery, SEO curation, the team and billing \| 120 per minute \| \| upload \| `POST /media/upload-url` and `POST /seo/backlinks/import` \| 60 per minute \| \| pipeline \| `POST /pipeline/runs`, `POST /pipeline/articles/{id}/requeue`, `POST /delivery/cms/push` and the seven paid SEO scans \| 10 per minute \| \| auth \| `POST /auth/device` and `POST /auth/device/token`, per IP address \| 60 per minute \| `POST /comments` adds three limits, named in `error.limit` on its 429: `reader` (5 per reader in 10 minutes, by `reader_ip`), `site` (300 per Site per hour) and `queue` (2,000 awaiting review). Every response carries `RateLimit-Limit`, `RateLimit-Remaining` and `RateLimit-Reset`. |
| 500 | - | `INTERNAL_ERROR`. Our fault. Quote the `request_id` from the body if you contact support. Safe to retry; send the same `Idempotency-Key` so a request that did succeed is not repeated. |
| 503 | - | `MAINTENANCE`. Writes are paused, platform wide or for your Site. Reads usually work and your published blog stays up. Retry after `Retry-After`. |

```bash
curl https://api.writavo.com/v1/formats/3f1b0c7a-0000-4000-8000-000000000001 \
  -H "Authorization: Bearer wv_sk_EXAMPLE0000000000000000000000000000"
```

### PATCH /formats/{id}

Change an article format

- **Operation id**: `updateFormat`
- **Scope**: `prompts:write`
- **Permission**: `prompts.manage`
- **Rate limit class**: write
- **Plan feature**: `none`
- **Spends credits**: no
- **Publishable key may call it**: no

Change a Site format's name, blueprint or whether it is active. The key cannot change. Articles
already written keep their structure; the change applies to articles written from now on.

A platform default is `409 CONFLICT` ("Platform formats cannot be changed. Create a Site format
with the same key to override it.").

**Parameters**

| Name | In | Type | Required | Description |
|---|---|---|---|---|
| `id` | path | uuid | yes |  |

**Request body** (required)

| Field | Type | Required | Description |
|---|---|---|---|
| `name` | string | no |  |
| `blueprint` | Blueprint | no | What a format tells the generator: the outline to follow and the structured blocks to fill. |
| `is_active` | boolean | no |  |
| `description` | string | null | no | `null` or `""` clears it. |
| `prompt_addendum` | string | null | no | `null` or `""` clears it. |

**Responses**

| Status | Data | Description |
|---|---|---|
| 200 | FormatDetail | The updated format. |
| 400 | - | `INVALID_REQUEST`. Bad JSON, an unusable query parameter value, or a missing required header. |
| 401 | - | `INVALID_API_KEY`, `API_KEY_REVOKED` or `API_KEY_EXPIRED`: no key, or an unusable one. It never distinguishes "no such key" from "wrong key". |
| 403 | - | `INSUFFICIENT_SCOPE`. The key lacks this operation's scope, or its creator's permissions no longer cover it. Another Site's object is always 404, never this. A 403 may also be: - `AGENT_ACCESS_DISABLED` - AI agent access is turned off. - `APPROVAL_DENIED` - a person denied this action. - `FORBIDDEN` (organisation, team, settings) - a change no scope allows, such as an AI agent changing the access of the person it acts for, anything about ownership, or a setting only a person may change. The message says what to do instead. |
| 404 | - | `NOT_FOUND`. No such object, or it belongs to another Site. **The two are deliberately indistinguishable and never return 403.** If the id is right, check you are using the right Site's key. |
| 409 | - | The request collides with the current state. Read `code`: - `SLUG_CONFLICT` - the slug is taken. - `IDEMPOTENCY_KEY_CONFLICT` / `IDEMPOTENCY_KEY_IN_FLIGHT` - key reused with another body, or still running. - `CONFLICT` - the object changed under you. Re-read and retry. - `APPROVAL_INVALID` - the `Writavo-Approval` is expired, used, unknown, for another request, or its effect changed. Retry without it for a new approval. - `PREREQUISITE_MISSING` / `FEATURE_UNAVAILABLE` (settings, delivery, SEO) - something is not set up (the message names what sets it up), or Writavo has it switched off (the message names the free alternative). Nothing was done or charged. |
| 422 | - | `VALIDATION_FAILED`. The values are not acceptable; `error.fields` maps each field to a message. Common causes: publishing without a title, slug or content; scheduling in the past; sending `status` on a create or update. |
| 429 | - | `RATE_LIMIT_EXCEEDED`. Back off and honour `Retry-After`. Limits are per key, per minute, by endpoint class. The two device sign-in endpoints take no key, so they are limited per IP address. \| Class \| Endpoints \| Limit \| \|---\|---\|---\| \| read \| every GET \| 600 per minute \| \| write \| POST, PATCH and DELETE on content, taxonomy, keys, webhooks, the content plan, formats and prompts, settings, delivery, SEO curation, the team and billing \| 120 per minute \| \| upload \| `POST /media/upload-url` and `POST /seo/backlinks/import` \| 60 per minute \| \| pipeline \| `POST /pipeline/runs`, `POST /pipeline/articles/{id}/requeue`, `POST /delivery/cms/push` and the seven paid SEO scans \| 10 per minute \| \| auth \| `POST /auth/device` and `POST /auth/device/token`, per IP address \| 60 per minute \| `POST /comments` adds three limits, named in `error.limit` on its 429: `reader` (5 per reader in 10 minutes, by `reader_ip`), `site` (300 per Site per hour) and `queue` (2,000 awaiting review). Every response carries `RateLimit-Limit`, `RateLimit-Remaining` and `RateLimit-Reset`. |
| 500 | - | `INTERNAL_ERROR`. Our fault. Quote the `request_id` from the body if you contact support. Safe to retry; send the same `Idempotency-Key` so a request that did succeed is not repeated. |
| 503 | - | `MAINTENANCE`. Writes are paused, platform wide or for your Site. Reads usually work and your published blog stays up. Retry after `Retry-After`. |

```bash
curl -X PATCH https://api.writavo.com/v1/formats/3f1b0c7a-0000-4000-8000-000000000001 \
  -H "Authorization: Bearer wv_sk_EXAMPLE0000000000000000000000000000" \
  -H "Content-Type: application/json" \
  -d '{}'
```

### DELETE /formats/{id}

Delete an article format

- **Operation id**: `deleteFormat`
- **Scope**: `prompts:write`
- **Permission**: `prompts.manage`
- **Rate limit class**: write
- **Plan feature**: `none`
- **Spends credits**: no
- **AI agent approval**: `format.delete`, always asked
- **Publishable key may call it**: no

Delete a Site format. Articles that use it keep their content and lose the format (their
`format_id` becomes null); plan items that name it fall back to the default. If the format
overrode a platform default, the default applies again. A platform default is `409 CONFLICT`.

**When an AI agent asks.** A key from a device or OAuth sign-in, or any request from an MCP
client, always gets `428 APPROVAL_REQUIRED` on the first try, whatever the organisation's
approvals switch says, because it changes how every future article is written and cannot be
undone. Give the person `error.approval.url`; once they approve, repeat the identical request
with `Writavo-Approval: <error.approval.id>`. The approval is void if what the request would do
changes before the retry. A key made in the dashboard is asked only when the request comes from
an MCP client.

**Parameters**

| Name | In | Type | Required | Description |
|---|---|---|---|---|
| `id` | path | uuid | yes |  |
| `Writavo-Approval` | header | uuid | no | The `error.approval.id` from a `428 APPROVAL_REQUIRED`, sent on the retry once a person has approved it. Only meaningful for a key approved for an AI agent; every other caller can leave it out. It is single use and bound to the key, to the exact request that asked for it (method, path and body) and to what that request would do: anything else, including a request whose effect changed since it was approved, is `409 APPROVAL_INVALID`. |

**Responses**

| Status | Data | Description |
|---|---|---|
| 204 | - | Deleted. No body. |
| 401 | - | `INVALID_API_KEY`, `API_KEY_REVOKED` or `API_KEY_EXPIRED`: no key, or an unusable one. It never distinguishes "no such key" from "wrong key". |
| 403 | - | `INSUFFICIENT_SCOPE`. The key lacks this operation's scope, or its creator's permissions no longer cover it. Another Site's object is always 404, never this. A 403 may also be: - `AGENT_ACCESS_DISABLED` - AI agent access is turned off. - `APPROVAL_DENIED` - a person denied this action. - `FORBIDDEN` (organisation, team, settings) - a change no scope allows, such as an AI agent changing the access of the person it acts for, anything about ownership, or a setting only a person may change. The message says what to do instead. |
| 404 | - | `NOT_FOUND`. No such object, or it belongs to another Site. **The two are deliberately indistinguishable and never return 403.** If the id is right, check you are using the right Site's key. |
| 409 | - | The request collides with the current state. Read `code`: - `SLUG_CONFLICT` - the slug is taken. - `IDEMPOTENCY_KEY_CONFLICT` / `IDEMPOTENCY_KEY_IN_FLIGHT` - key reused with another body, or still running. - `CONFLICT` - the object changed under you. Re-read and retry. - `APPROVAL_INVALID` - the `Writavo-Approval` is expired, used, unknown, for another request, or its effect changed. Retry without it for a new approval. - `PREREQUISITE_MISSING` / `FEATURE_UNAVAILABLE` (settings, delivery, SEO) - something is not set up (the message names what sets it up), or Writavo has it switched off (the message names the free alternative). Nothing was done or charged. |
| 428 | - | `APPROVAL_REQUIRED` or `APPROVAL_PENDING`. Only ever returned to an AI agent: a key from a device or OAuth sign-in, or any request an MCP client makes (it carries `Writavo-Mcp-Tool`) whatever key it holds. For an operation marked `x-writavo-approval-always` it is returned whatever the organisation's approvals switch says, and for the six content actions while approvals are on. Nothing has happened yet. Show `error.approval.url` to a person; once they approve, repeat the identical request with `Writavo-Approval: <error.approval.id>`. Do not change the body or the path: the approval is bound to the exact request, and a different one is `409 APPROVAL_INVALID`. |
| 429 | - | `RATE_LIMIT_EXCEEDED`. Back off and honour `Retry-After`. Limits are per key, per minute, by endpoint class. The two device sign-in endpoints take no key, so they are limited per IP address. \| Class \| Endpoints \| Limit \| \|---\|---\|---\| \| read \| every GET \| 600 per minute \| \| write \| POST, PATCH and DELETE on content, taxonomy, keys, webhooks, the content plan, formats and prompts, settings, delivery, SEO curation, the team and billing \| 120 per minute \| \| upload \| `POST /media/upload-url` and `POST /seo/backlinks/import` \| 60 per minute \| \| pipeline \| `POST /pipeline/runs`, `POST /pipeline/articles/{id}/requeue`, `POST /delivery/cms/push` and the seven paid SEO scans \| 10 per minute \| \| auth \| `POST /auth/device` and `POST /auth/device/token`, per IP address \| 60 per minute \| `POST /comments` adds three limits, named in `error.limit` on its 429: `reader` (5 per reader in 10 minutes, by `reader_ip`), `site` (300 per Site per hour) and `queue` (2,000 awaiting review). Every response carries `RateLimit-Limit`, `RateLimit-Remaining` and `RateLimit-Reset`. |
| 500 | - | `INTERNAL_ERROR`. Our fault. Quote the `request_id` from the body if you contact support. Safe to retry; send the same `Idempotency-Key` so a request that did succeed is not repeated. |
| 503 | - | `MAINTENANCE`. Writes are paused, platform wide or for your Site. Reads usually work and your published blog stays up. Retry after `Retry-After`. |

```bash
curl -X DELETE https://api.writavo.com/v1/formats/3f1b0c7a-0000-4000-8000-000000000001 \
  -H "Authorization: Bearer wv_sk_EXAMPLE0000000000000000000000000000"
```
