You pay based on the number of tokens processed. For a typical document like an invoice:
Cost varies based on:
A multi-page legal contract will cost more than a simple receipt. You can check usage costs in your dashboard after processing.
No credit card required to start
We can build custom features, integrations, or deploy on your infrastructure. Every business has unique needs - let's discuss yours.
SOC2, GDPR, HIPAA compliance frameworks. On-premise deployment available.
Real-time confidence scoring, automated validation, custom workflows, and dedicated infrastructure.
Processing thousands of documents monthly? We'll work out pricing that makes sense for you.
30-minute demo and discussion
Built by engineers who got frustrated with existing document extraction tools
Our augmented text processing uses position markers to extract the right "John Smith" when there are five of them in the document. Most tools just guess.
Try extracting "supplier name" from an invoice that also has customer name, company name, and delivery contact. You'll see why this matters.
Extract complex data like "list of line items, each with product name, quantity, and price" or "employment history with dates and duties for each job."
Most competitors force you to flatten everything or make multiple API calls. We handle nested objects and arrays natively.
K-Score validation gives you a 0-100 confidence rating per field. A score of 100 means we tested it 100 times with high temperature and got consistent results.
Build business logic based on confidence: auto-process high-score extractions, flag low-score ones for human review.
Every time you validate or correct an extraction, it becomes training data. Our one-out testing continuously validates and improves accuracy.
Unlike other tools where you're stuck with whatever accuracy you get, DeepTagger gets better the more you use it.
You can check usage costs in your dashboard after processing. We're working on a pre-processing cost estimator. Generally, simple invoices/receipts cost $0.05-0.10, while complex multi-page contracts might be $0.20-0.50 depending on how much data you're extracting.
We use the same token counting as OpenAI (since we use their models under the hood). Roughly speaking: your document content + your training examples + your output schema = total tokens. A typical invoice is 1,000-2,000 tokens of input.
Yes. If you're processing thousands of documents per month, book a call with us. We can work out pricing that makes sense for your volume. We'd rather have you as a long-term customer than nickel-and-dime you on per-document pricing.
If you're not getting good results, reach out to us first. Often it's a matter of adding better training examples or adjusting your schema. But yes, if something's genuinely broken or not working as described, we'll refund you. We're not trying to trap you.
$5 gives you roughly 50-100 document extractions (depending on document complexity). That's enough to properly test DeepTagger with your real documents and see if it works for your use case. No time limit on using those credits.
Start with $5 in free credits. See if DeepTagger works for your documents. No credit card required.
Questions? Email us at hello@deeptagger.com
Pricing updated October 2025. Cost per document varies based on complexity, training examples, and output structure. Enterprise pricing and custom deployments available - just ask.