REST API Reference

Start the server with adx serve (default: http://localhost:8000).

Files

POST /v1/files

Upload a document.

curl -X POST http://localhost:8000/v1/files -F "file=@document.pdf"

Response:

{"id": "doc_abc123", "file_name": "document.pdf", "file_type": "pdf", "status": "ready"}

GET /v1/files/

Get file metadata.

GET /v1/files

List all uploaded files.

Inspection

GET /v1/files/{id}/profile

Profile the document — metadata, type detection, recommended tools.

GET /v1/files/{id}/structure

Document structure — sections, tables, page outline.

POST /v1/files/{id}/search

Search the document.

{"query": "payment terms", "max_results": 20}

GET /v1/files/{id}/pages/

Get a specific page's text blocks and tables.

GET /v1/files/{id}/tables

List all tables in the document.

GET /v1/files/{id}/tables/

Get a specific table's contents.

Spreadsheet Tools

GET /v1/files/{id}/sheets

List sheets with metadata.

POST /v1/files/{id}/ranges/read

Read a cell range.

{"sheet": "Summary", "range": "A1:D10", "include_formulas": true}

POST /v1/files/{id}/cells/find

Find cells matching a query.

{"query": "Revenue", "sheet": "Summary"}

POST /v1/files/{id}/formulas/inspect

Inspect a formula's dependencies.

{"sheet": "Summary", "cell": "B10"}

Extraction

POST /v1/files/{id}/extract

Extract fields using a schema.

{"schema": "invoice"}

Or with a custom schema:

{
  "schema": {
    "name": "custom",
    "fields": [
      {"name": "po_number", "type": "string", "required": true}
    ]
  }
}

GET /v1/extractions/

Retrieve a previous extraction result.

Validation

POST /v1/files/{id}/validate

Validate an extraction.

{"extraction_id": "ext_abc123"}

Citations

GET /v1/files/{id}/citations

Get all citations for a document.

Export

POST /v1/files/{id}/export

Export document content.

{"format": "markdown"}

Supported formats: json, markdown, csv.

Health

GET /health

{"status": "healthy", "version": "0.1.0"}