Documents

Endpoints

Method	Path	Scope	Description
GET	/documents	documents:read	List documents
POST	/documents	documents:write	Create a document
GET	/documents/:id	documents:read	Get a document with content
PATCH	/documents/:id	documents:write	Update a document
DELETE	/documents/:id	documents:write	Soft-delete a document

List Documents

bash

curl "https://my.clarife.app/api/v1/documents?limit=10&offset=0" \
  -H "Authorization: Bearer clrf_your_key"

Query parameters:

Param	Type	Default	Description
limit	integer	50	Results per page (1-100)
offset	integer	0	Number of results to skip
sort	string	updated_at	Sort field: updated_at, created_at, title
order	string	desc	Sort order: asc or desc
project_id	uuid	--	Filter by project

Response:

json

{
  "data": [
    {
      "id": "d1a2b3c4-5678-90ab-cdef-1234567890ab",
      "title": "Getting Started Guide",
      "description": "Onboarding tutorial for new users",
      "updated_at": "2026-03-20T14:00:00Z",
      "created_at": "2026-03-15T10:30:00Z",
      "block_count": 12,
      "screenshot_count": 5,
      "word_count": 450,
      "visibility": "private"
    }
  ],
  "total": 1,
  "limit": 10,
  "offset": 0
}

Create a Document

bash

curl -X POST https://my.clarife.app/api/v1/documents \
  -H "Authorization: Bearer clrf_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "API Setup Guide",
    "description": "How to connect via the clarife API",
    "project_id": "p1a2b3c4-5678-90ab-cdef-1234567890ab"
  }'

Body parameters:

Field	Type	Required	Description
title	string	No	Document title (max 500 chars, defaults to "Untitled")
description	string	No	Short description (max 2000 chars)
content	object	No	Document content (see Content Format below)
project_id	uuid	No	Assign to a project
template	string	No	Template identifier
visibility	string	No	private or workspace (requires X-Workspace-Id header)

Response (201 Created):

json

{
  "data": {
    "id": "d2b3c4d5-6789-01ab-cdef-2345678901ab",
    "title": "API Setup Guide",
    "created_at": "2026-03-28T12:00:00Z"
  }
}

Get a Document

bash

curl https://my.clarife.app/api/v1/documents/d1a2b3c4-5678-90ab-cdef-1234567890ab \
  -H "Authorization: Bearer clrf_your_key"

Returns the full document including content with all blocks.

Response:

json

{
  "data": {
    "id": "d1a2b3c4-...",
    "title": "Getting Started Guide",
    "description": "Onboarding tutorial",
    "content": {
      "meta": { "version": 1, "locale": "en" },
      "blocks": [
        { "type": "heading", "id": "h1-title", "sort_order": 0, "level": 1, "content": "Getting Started" },
        { "type": "text", "id": "t1-intro", "sort_order": 1, "content": "Welcome to the guide." },
        { "type": "screenshot", "id": "ss1-dashboard", "sort_order": 2, "media_id": "m1234567-...", "caption": "Dashboard view", "alignment": "center" }
      ]
    },
    "version": 4,
    "block_count": 3,
    "screenshot_count": 1,
    "word_count": 12,
    "visibility": "private",
    "template": null,
    "branding_id": null,
    "updated_at": "2026-03-20T14:00:00Z",
    "created_at": "2026-03-15T10:30:00Z"
  }
}

The top-level version field is the document's version number, incremented on every content change - pass it as expected_version when updating (see below).

Update a Document

bash

curl -X PATCH https://my.clarife.app/api/v1/documents/d1a2b3c4-5678-90ab-cdef-1234567890ab \
  -H "Authorization: Bearer clrf_your_key" \
  -H "Content-Type: application/json" \
  -d '{ "title": "Updated Title" }'

All body fields are optional. Only provided fields are updated.

Field	Type	Description
title	string	New title (max 500 chars)
description	string	New description (max 2000 chars)
content	object	Replace document content
expected_version	integer	Optional concurrency guard - the version value from GET /documents/:id (mismatch = 409 CONFLICT)

⚠️ Warning: When updating content, you must send the full content - the blocks array replaces the existing one. First read the current content with GET /documents/:id, modify it, and send it all back. The only exception are block_annotations and block_enhance_settings - you can omit them and the server keeps the existing values (see Annotations and screenshot enhance settings below).

Optimistic concurrency (optional): pass the version value from the GET /documents/:id response as expected_version. The update then applies only if the document has not changed in the meantime - otherwise the API returns 409 with code CONFLICT and the current version number in the message. On a conflict, re-read the document, re-apply your changes to the fresh content, and retry. We recommend sending expected_version with every content update - it protects against overwriting changes saved concurrently in the editor, by another integration, or by an MCP agent.

Response (200 OK):

json

{
  "data": {
    "id": "d1a2b3c4-...",
    "title": "Updated Title",
    "version": 5,
    "updated_at": "2026-03-28T13:00:00Z"
  }
}

Delete a Document

bash

curl -X DELETE https://my.clarife.app/api/v1/documents/d1a2b3c4-5678-90ab-cdef-1234567890ab \
  -H "Authorization: Bearer clrf_your_key"

Documents are soft-deleted (moved to trash). They are permanently removed after 30 days.

Response (204 No Content): Empty body.

Content Format

Document content follows a flat block-based structure. Each block must include these common fields:

Field	Type	Required	Description
type	string	Yes	Block type (see types below)
id	string	Yes	Unique ID within the document (e.g. "h1-title", "ss3-result")
sort_order	integer	Yes	Display order starting from 0

heading

json

{
  "type": "heading",
  "id": "h1-title",
  "sort_order": 0,
  "level": 1,
  "content": "My Tutorial Title"
}

level: 1, 2, or 3 (H1, H2, H3)
content: plain text

text

json

{
  "type": "text",
  "id": "t1-intro",
  "sort_order": 1,
  "content": "Paragraph text here.\nUse line breaks for lists:\n• Item one\n• Item two"
}

content: plain text or basic HTML (<b>, <i>, <a>, <br>)

screenshot / image

json

{
  "type": "screenshot",
  "id": "ss1-dashboard",
  "sort_order": 2,
  "media_id": "77b4185f-bb13-47ea-97a4-6b0d88289df1",
  "caption": "Dashboard overview",
  "alignment": "center"
}

media_id (UUID, required): references an uploaded media file - see the Media API for the upload workflow
caption (string, optional): displayed below the image
alignment: "left", "center", or "right"
width (integer, optional): display width in pixels (omit for original size)

Use "image" type for non-screenshot visuals (diagrams, illustrations). Same schema as screenshot.

divider

json

{ "type": "divider", "id": "d1", "sort_order": 5 }

No additional fields.

code

json

{
  "type": "code",
  "id": "c1-example",
  "sort_order": 6,
  "code": "npm install clarife",
  "language": "bash"
}

code (required): the code string
language: syntax highlighting (e.g. "python", "javascript", "bash", "sql")

table

json

{
  "type": "table",
  "id": "tbl1",
  "sort_order": 7,
  "has_header": true,
  "columns": [
    { "label": "Name" },
    { "label": "Role" }
  ],
  "rows": [
    ["Alice", "Developer"],
    ["Bob", "Designer"]
  ]
}

has_header (boolean, required): whether the column labels render as a header row
columns (required): one entry per column; label is the header text, optional width (px) and align ("left" | "center" | "right")
rows: 2D array of cell strings - data only, headers live in columns

Annotations and screenshot enhance settings

Besides meta and blocks, content can carry two optional objects created in the clarife editor:

block_annotations: annotation layers (arrows, boxes, blur, etc.) keyed by screenshot block id
block_enhance_settings: screenshot frame and background styling, also keyed by block id

Both fields are read-only through the API. When updating a document you can omit them - the server keeps the existing values. You can also send them back exactly as returned by GET /documents/:id. Any new or changed value is rejected with 400 and code ANNOTATIONS_EDITOR_ONLY.

ℹ️ Why: public share and embed pages (and AI videos) show images baked in the clarife editor. The API does not generate baked images, so annotations written through the API would never be visible to readers - and blur or pixelation would not really hide the data underneath. That is why the API rejects these fields instead of accepting content readers would never see. To point at something in a screenshot, describe it in the step text - the document owner can add arrows, step markers, or blur in the editor.

Complete Example

Creating a document with text and a screenshot:

bash

curl -X POST https://my.clarife.app/api/v1/documents \
  -H "Authorization: Bearer clrf_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Setup Guide",
    "content": {
      "meta": { "version": 1, "locale": "en" },
      "blocks": [
        { "type": "heading", "id": "h1", "sort_order": 0, "level": 1, "content": "Setup Guide" },
        { "type": "text", "id": "t1", "sort_order": 1, "content": "Follow these steps to get started." },
        { "type": "screenshot", "id": "ss1", "sort_order": 2, "media_id": "77b4185f-...", "caption": "Settings screen", "alignment": "center" },
        { "type": "divider", "id": "d1", "sort_order": 3 },
        { "type": "text", "id": "t2", "sort_order": 4, "content": "You are all set!" }
      ]
    }
  }'

ℹ️ Info: To add images, you must first upload them via the Media API. The 4-step workflow is: create presigned URL → upload file → confirm → add screenshot block with media_id to the document.