Thinking Mode is Veryfi's agentic data extraction. Add the thinking parameter to a document request, and AI agents review the extraction, catch mistakes and correct them before the results come back. It currently works with receipts and invoices and needs to be enabled for your account first.
What's the difference between light and deep?
The value you send decides how many agents join the review.
Value | What happens | Use it for |
Not sent (default) | Standard real-time pipeline, under 3 seconds per document | Live flows where someone is waiting on the result |
| A small team of agents reviews the extraction and corrects mistakes | A quick second look on everyday documents |
| A bigger team of agents runs the most thorough review | Hard or high-stakes documents |
Which documents does Thinking Mode support?
Thinking Mode currently supports receipts and invoices, processed through the documents endpoint: POST https://api.veryfi.com/api/v8/partner/documents.
When should I use Thinking Mode?
Use it when a few extra seconds are worth higher accuracy, for example:
Invoices that post straight to your ERP or accounting system with no human review
Expense reports and reimbursements, where a wrong total means a wrong payout
Back-office and batch jobs, where nobody is waiting on the result
Skip it when you need an instant answer, such as a customer capturing a receipt in your app. Thinking Mode adds a few seconds per document, so the standard pipeline is the better fit there.
How do I get access?
To enable Thinking Mode for your account, message us in the Slack channel your organization shares with Veryfi, or email support@veryfi.com.
How do I process a document with Thinking Mode?
Add the thinking parameter to your POST request, set to light or deep. The rest of the request stays the same.
curl -X POST https://api.veryfi.com/api/v8/partner/documents \
-H "Content-Type: application/json" \
-H "Client-Id: <client_id>" \
-H "Authorization: apikey <username>:<api_key>" \
-d '{
"file_url": "https://example.com/invoice.pdf",
"thinking": "light"
}'
If you don't send thinking, nothing changes: the document goes through the standard real-time pipeline. For the full request reference, see the document processing guide.
Where do Thinking Mode's agents run?
Thinking Mode's agents run on Veryfi's own isolated infrastructure, just like our extraction models.