# Make scanned PDFs readable and structured Canonical page: https://districtproof.io/services/scanned-document-ocr Optical character recognition turns a picture of text into machine-readable text. Accessibility work then checks recognition accuracy, structure, reading order, and information that OCR cannot describe. ## What this service covers - Text recognition and correction - Headings and reading sequence - Tables and visual information - Document structure and review ## Start with the task and the source A scanned page starts as an image. Without recognized text, searching, selecting, or reading it through many assistive technologies can be difficult. OCR creates a text layer, but it can misread characters, rearrange columns, or miss handwritten content. ## What needs human attention Review quality depends on the original scan. Low contrast, skewed pages, small type, stamps, and handwritten notes can require additional attention. Supply a clean scan or an editable original when available. A sample review helps determine what can be recovered reliably. ## Make the next update easier After OCR, the work continues. Add appropriate structure, check reading order, identify headings and tables, and describe meaningful non-text content. A successful recognition pass does not mean that a document is ready for every reader. ## Scope the work before setting the schedule Agree on the inputs, document versions, accepted formats, complexity, review method, and output before work starts. Flag any required publishing dates, confidential material, or dependencies on a website vendor. A representative sample helps avoid assumptions about the rest of the backlog. ## Pricing and next steps This service is part of [DistrictProof Docs](https://districtproof.io/solutions/docs). [Request a fixed quote](https://districtproof.io/pricing) after a sample review or begin with your public website inventory. For the regulatory context, read the [ADA Title II school district deadline guide](https://districtproof.io/resources/ada-title-ii-deadline-school-districts). Your district and counsel determine the applicable obligations and exceptions. ## Find the best available original before recognizing text A scan is sometimes the only surviving record, but it is not always the best available source. An office may still have the original Word file, a clean export from a reporting system, or a higher-quality scan. Ask the content owner before investing in repair. A better original can reduce recognition errors and preserve information that a compressed copy has lost. For image-only files, inspect orientation, page completeness, contrast, skew, and small print. A scan with missing edges or unreadable handwriting needs a content decision. OCR cannot reliably reconstruct information that is not visible. We identify those limits during the sample review rather than allowing uncertain text to pass unnoticed. Illustrative district scenario ### A scanned school menu with a visual grid Recognition may recover meal names while mixing the days of the week. A useful result must preserve which meal belongs to which day and how any notes apply. The review considers the relationships as well as the recognized words. The original visual page remains a reference when checking the result. ## Recognition quality needs content-level attention Numbers, dates, names, and identifiers deserve particular care. A single incorrect character in a contact number or application instruction can change what a family does next. Similar-looking characters and merged words are common review targets. The district should resolve unclear originals rather than asking a remediation team to infer an official fact. Tables, multi-column pages, stamps, signatures, and handwritten notes can require different handling. Decide whether a mark is meaningful, decorative, or a sign that the source needs clarification. Record any limitation that affects the information available to a reader. ## Move from recognized words to usable structure After recognition, organize the document into headings, paragraphs, lists, and tables that reflect its meaning. Check whether text is being read in the intended order. Meaningful graphics still need an appropriate text treatment. Searchable text helps, but the final experience also depends on structure and context. For recurring material, use the review to identify a better publishing path. A frequently updated form or notice may be more sustainable as an editable template or a web page. A historic scan may remain a fixed document. Those are separate content and records decisions that the district can make with the right people involved. ## How scanned work affects a fixed quote Page count is only the starting point. Scan quality, density, language, handwriting, tables, and the availability of an original influence the effort. Include the difficult pages in the sample so the quote covers the actual work. State whether the deliverable is a remediated PDF, a corrected source, or an agreed replacement format. For a mixed backlog, [DistrictProof Docs](https://districtproof.io/solutions/docs) can separate scanned material from simpler text documents. If the same issues recur, [accessible templates](https://districtproof.io/services/accessible-templates) help staff avoid creating a new scan-based workflow for every edition. ## Frequently asked questions ### What do you need to scope the work? A representative sample, the best available source files, the intended use, and your required publishing schedule. Website projects also need an agreed list of pages, templates, and tasks. ### How is this priced? We review complexity and scope before providing a fixed quote. Page counts alone do not describe the effort required for scans, tables, forms, or interactive components. ### Will an automated checker be enough? A checker can identify some issues, but a human review is needed for meaning, order, descriptions, and use. Evidence should identify both the tools and the manual coverage. ### What will we receive? Your agreement defines the returned files or issue log, review coverage, delivery schedule, and evidence record. Publication verification should reference the live version.