What we OCR
- Scanned PDFs (single-page and multi-page)
- Photos of documents (JPG, PNG, HEIC, raw from phones)
- Faxes (often low-resolution and noisy)
- Whiteboard photos
- Receipts and invoices
- ID cards and passports (for verification workflows, with appropriate redaction)
- Handwritten notes and forms
- Multi-language documents (any combination of 100+ languages)
- Historical documents (old fonts, faded ink, special characters)
- Tables and structured layouts
Accuracy and quality
OCR accuracy varies dramatically based on the input.
- Clean typewritten text — 99%+ accuracy, suitable for direct use
- Standard scanned documents — 95-98% accuracy, human review recommended
- Poor scans, faded, low DPI — 85-95% accuracy, human review essential
- Handwritten text — 70-90% accuracy depending on legibility, always human-reviewed
- Special fonts, non-Latin scripts, mixed layouts — varies; we will quote a sample first
Our human review pass takes the OCR output and corrects any errors, especially around tables, headers, special characters, and proper nouns. The result is a clean, accurate, ready-to-use document.
Output formats
- Microsoft Word (.docx) — most common, preserves layout
- Microsoft Excel (.xlsx) — for tabular data, with formulas and formatting where detectable
- Plain text (.txt) — for simple text-only output
- Searchable PDF — original PDF with a text layer added (so text is selectable, searchable, accessibility-compliant)
- JSON / XML — for structured data extraction with positional information
- CSV — for tabular data in spreadsheet-friendly form
Industry use cases
- Legal — discovery documents, contracts, court bundles. Searchable PDFs for e-discovery
- Medical — patient records, clinical notes, lab reports. HIPAA-compliant handling
- Finance — invoices, receipts, bank statements, contracts. Multi-format output
- Real estate — leases, title documents, inspection reports
- Education — student records, transcripts, certificates
- Government — archives, FOI requests, public records digitisation
- Research — academic papers, historical documents, manuscripts
Pricing
OCR is priced per page based on scan quality.
- Clean scans — £10 per page
- Standard scans — £15 per page
- Poor scans or handwriting — £20 per page
Bulk discounts apply for 10+ pages. Multi-language documents may incur a small surcharge depending on the script.
Turnaround
Standard delivery: 48 hours. Most small OCR jobs (under 50 pages) are delivered within 24 hours.
Express delivery: 12 hours, +£20 per order.
For very large jobs (1,000+ pages), we can parallelize across multiple editors and deliver in 2-3 business days.