Overview
All Aifano processing endpoints return structured JSON with a consistent format. The core building blocks are chunks (logical sections of content) and blocks (individual elements like text, tables, and figures).Parse Response
Top-Level Fields
| Field | Type | Description |
|---|---|---|
job_id | string | Unique identifier for this processing job |
duration | number | Processing time in seconds |
usage | object | Credit and page usage information |
result | object | The processing result containing chunks |
Chunks
Chunks are logical sections of the document. Each chunk contains:| Field | Type | Description |
|---|---|---|
content | string | Markdown-formatted content of the section |
embed | string | Embedding-optimized plain text (ideal for RAG) |
blocks | array | Individual elements with bounding boxes |
Blocks
Blocks are the atomic elements within a chunk. Each block has:| Field | Type | Description |
|---|---|---|
type | string | The block type (see table below) |
content | string | The text or data content |
bbox | object | Bounding box coordinates |
Block Types
| Type | Description |
|---|---|
Title | Document or section title |
Section Header | Sub-section heading |
Text | Body text paragraph |
Table | Tabular data (Markdown format) |
Figure | Image or chart |
List Item | Bulleted or numbered list item |
Header | Page header |
Footer | Page footer |
Page Number | Page number |
Key Value | Key-value pair |
Comment | Annotation or comment |
Signature | Signature block |
Bounding Boxes
All coordinates are normalized (0.0 to 1.0) relative to the page dimensions:| Field | Type | Description |
|---|---|---|
left | number | Left edge (0.0 = left margin, 1.0 = right margin) |
top | number | Top edge (0.0 = top, 1.0 = bottom) |
width | number | Width as fraction of page width |
height | number | Height as fraction of page height |
page | integer | 1-based page number |
Usage Object
| Field | Type | Description |
|---|---|---|
num_pages | integer | Number of pages processed |
credits | integer | Credits consumed for this operation |
Extract Response
Extract responses include the parsed data plus extracted fields:result matches the JSON schema you provided in the request.