ComPDF AI Overview

ComPDF AI provides Intelligent Document Processing (IDP) APIs under the POST /v2/process/idp/* family. These guides explain concepts, request modes, parsing configuration, and extraction schemas. For full field-level API details, see the corresponding API Reference.

Capabilities

Capability	Path	Purpose
Document Parsing	idp/documentParsing	Parse document structure and produce JSON / tagged Markdown with bbox references
Document Extraction	idp/documentExtract	Extract key fields and table data; supports layout and vision modes

Core concepts

• Task — each parsing or extraction call creates a task and returns task metadata. Sync mode returns the result directly; async and presigned flows require status polling.
• Extraction schema — describes the keys and table headers to extract. Both layout and vision modes take one fixed schema via extract_fields. The schema shape is { name, keys, tableHeaders }.
• BBox grounding — layout mode can return page and bbox references so you can map results back to the source document.

Recommended path

1. Prerequisites
2. First request: sync extraction
3. Request modes
4. Document parsing guide
5. Document extraction guide
6. FAQ & Billing