Legal & claims processing

Handwriting OCR is an AI service that turns handwritten documents into accurate, searchable, editable text and structured data. Unlike traditional OCR, which was built for printed text, it is built specifically for handwriting, including cursive, faded ink, and many languages, and it can translate non-English documents into English in the same step. Upload a scan or photo and you get back clean text you can search, edit, and export to Word, Markdown, plain text, or a spreadsheet.

Handwriting OCR was founded in London in 2023, dedicated to applying modern AI to read the hardest handwritten documents that traditional OCR cannot handle. We are a small, independent UK team, the people who build the product also handle support, and we never use your documents to train models or share them with anyone.

Yes. You can automate uploads and retrieval over the API with webhooks to process documents continuously as part of your own workflow. For very large backlogs, typically tens of thousands of pages, we offer managed processing: you send us the files, we process them with a configuration tuned to your document types, and deliver the results in days. It accepts scanned PDFs and common image formats (JPG, PNG, TIFF) without preprocessing.

Yes. On the Pro and Business plans, custom extractors let you define the fields you want, such as claimant, policy number, dates of loss, or amounts, and pull them off every form in a batch into a spreadsheet (XLSX, CSV, or JSON). Table extraction does the same for tabular records like schedules of loss. Extractors are callable over the API using an extractor ID, so they slot into an automated pipeline.

Yes. Handwriting OCR is built for mixed-content pages where printed text and handwriting appear together, which is common in legal work: claim forms with handwritten answers, contracts with handwritten amendments, or typed reports with handwritten annotations. It recognizes both and preserves the structure of the page, so you can see which portions were original and which were added by hand.

Your documents remain private and are processed only to deliver results to you. They are encrypted in transit and at rest, never used to train AI models, not shared with third parties, and not retained longer than necessary (you control retention, with auto-delete configurable from 15 minutes up to 14 days). Confidentiality is built into the service design as a fundamental principle, not an optional feature.

Yes, on the Business plan. It includes a GDPR Data Processing Agreement and a security documentation packet for procurement reviews. EU customers can choose EU-only data residency, and region-specific residency can be arranged for enterprise as part of a custom package. We also accept purchase orders for institutional and enterprise customers. We do not currently hold SOC 2 or ISO 27001, so if your firm has specific requirements, get in touch and we can talk through exactly how processing works.

Accuracy depends on the handwriting and the scan. Handwriting OCR is built for cursive and rushed hands and reads real-world documents far better than OCR designed for print, and clear scans process very well. Severely degraded pages, very faded ink, and stylized signatures are harder and need more review. The best guide for a given source is to test a representative sample, which is also how a managed-processing job is scoped.

No, and it is not designed to. Handwriting OCR accelerates the mechanical work of converting handwriting to text and structured data, but legal documents still require professional review. Context-dependent abbreviations, legal terminology, and figures that must reconcile need human verification. The tool removes the transcription bottleneck so your people spend their time on analysis and judgment rather than data entry.