Uploading Documents

NetraOCR accepts documents in three ways. All of them are available both in the console and over the API.

Supported formats & limits

Property	Value
File types	PDF, PNG, JPG/JPEG, TIFF, BMP, WebP
Max single file	50 MB
Max bulk total	600 MB
PDF pages	Multi-page PDFs are fully supported

Page limits

Per-file page caps may apply depending on your plan (and, on Business plans, on the specific template). The console shows the effective limit before you upload.

1. Single upload (synchronous)

The simplest flow — upload one file and get the result back in the same response.

In the console: Dashboard → Upload → choose a file → select a template → optionally toggle integrity → Submit. The result appears in seconds.

Over the API: POST /ocr/upload with multipart/form-data. See OCR endpoints.

Good for: interactive use, one-off documents, low-latency integrations.

2. Base64 upload (synchronous)

Identical to single upload, but the file is sent as a Base64 string inside a JSON body instead of multipart form data. Ideal when multipart is inconvenient — for example calling from a serverless function or a system that only speaks JSON.

Over the API: POST /ocr/upload-base64. See OCR endpoints.

3. Bulk upload (asynchronous)

Submit many files at once — a whole folder of invoices, for example. Bulk uploads are processed in the background and grouped into a batch.

The recommended flow is two steps:

1. Analyze (POST /ocr/bulk-analyze) — a pre-flight check that validates every file (type, size, page count) and returns an accurate cost estimate. It does not create jobs.
2. Upload (POST /ocr/bulk-upload) — submits the batch for processing and returns a batch_id. Watch the batch, or poll individual jobs, to collect results.

In the console: Dashboard → Upload → select multiple files / a folder. The UI runs the analysis automatically and shows you the cost before you confirm.

See OCR endpoints, Batches, and Jobs & Results for the API details.

Choosing a template at upload time

Every OCR upload requires a selected template. Over the API, pass its template_id so the result is returned as structured data keyed to your fields:

• Pass a global template ID (e.g. gl1) → structured data using a built-in schema.
• Pass a user template ID → structured data using your custom schema.

Learn how to create templates in Templates.

Enabling document integrity

Set the integrity option on upload to additionally screen the document for tampering. The check runs in the background and is retrieved separately — see Document Integrity.

What you get back

Field	Description
job_id	Unique ID for this processing job
structured_data	Typed fields from the selected template
page_count	Number of pages processed
document_type	Detected document type (e.g. invoice)
processing_time_ms	How long processing took
coins_consumed	Usage charged for this job

Next: review and export results in Results & Exports.