Starter
For focused document testing
- All 6 standard document types
- Custom document uploads
- Synthetic PDF documents
- JSON and CSV labels

Real-world documents.
Entirely fictional data.
Labeled datasets for AI training and OCR testing. Start with a document type or bring your own sample.
01 / Precision by design
Export PDFs with JSON or CSV labels, or choose standalone JSON, JSONL or CSV. Standard document totals are checked.
Generated test data, not real personal records. Every PDF is clearly watermarked as synthetic test data.
Same settings and seed, same dataset — every time. Perfect for regression tests and benchmarks.
Upload a PDF, image or text sample. Turn its fields into fictional records and labeled PDFs in a simplified layout — without copying its original values.
02 / The pipeline
Pick a document type or upload your sample
Choose your batch size and export format
Download synthetic documents and labels
Plans & pricing
Four tiers for your next AI project. Custom document uploads included across every tier.
Your own documents. New fictional data.
For focused document testing
For ongoing AI dataset projects
For broader document test coverage
For comprehensive AI evaluation
Upload a PDF, image or text sample. AI identifies its fields, then new fictional values become labeled documents in a simplified layout. Your source file is never included in the download.
Current workspace limit: 1,000 documents per batch. Samples: up to 10 MB; PDFs up to 10 pages. Custom totals are not mathematically checked.
Plan pricing shown in USD. No free tier. Feature lists highlight current workspace capabilities, not enforced plan restrictions. All tiers currently open the same workspace; subscriptions are not available for purchase yet.
Standard types or your own sample. Always fictional data.
Create custom documents