The document processing workflow

Availability: Arranged through sales

The API organizes work the way Trace does: a project contains documents, each document yields one or more legal descriptions, and each legal description carries the survey calls that define its geometry. Exports exist at two levels: per legal description, and consolidated per project.

The workflow

1. Create a project. POST /projects creates the container. Instant.
2. Add a document, two ways.
   • From text: POST /projects/{project-id}/documents/from-legal-description processes raw description text you already have. The pipeline extracts the calls, structures them, and generates the geometry.
   • From a file: POST /projects/{project-id}/documents/from-file uploads a document. OCR runs where needed, the pipeline identifies and extracts the legal descriptions, then plots them. Supported formats: PDF, JPEG, PNG, TIFF. Processing takes 1 to 4 minutes depending on the document.

   Each successful document processing consumes 1 credit.

3. Include the document in project output. PUT /projects/{project-id}/documents/{document-id}/include marks it for the consolidated project export. Documents can also be auto-included with a flag at upload.
4. Retrieve results.
   • GET /projects/{project-id} returns the full project, including the consolidated GeoJSON of every included legal description.
   • GET /projects/{project-id}/legal-descriptions/{legal-description-id} returns one description: its calls and its generated geometry.
   • GET /projects lists your projects by name and id.

Working notes

• Text input processes faster than file upload; send text when you already have clean description text.
• Scan quality decides extraction quality. 300 DPI or better for file uploads.
• Check response status before the next step, and handle credit consumption in your error paths so a retry does not double-process.

Related

• What the API returns
• Authentication
• Limitations