A11yReady
AI-Powered Document Accessibility Platform
Platform Screenshots

Branded upload landing with drag-and-drop support, document type auto-detection, and per-department logo selection

Reviewer compares the original PDF and the accessible HTML side by side, then approves and exports as HTML+PDF, HTML only, or PDF only

Quality gate surfaces specific structural issues — missing content, light text contrast, table count mismatches — with one-click fix workflow

Reviewer types a plain-English fix instruction or clicks a suggestion; the deterministic directives loop turns each correction into a safe per-file change applied on the next run

Searchable conversion history with per-document quality score, review status, and full audit trail across all uploads

Admin view: documents converted, success rate, avg quality score, avg extraction time, estimated time saved, estimated cost saved (manual labor vs. AI compute), and quality-trend chart over the last 30 days
Overview
Production AI platform clearing Multnomah County's public-PDF backlog ahead of the ADA Title II digital-accessibility deadline (extended to April 2027). Documents flow through LangGraph orchestration with per-cluster model routing on Vertex AI: Gemini 2.5 Flash for vision on most documents, Claude Sonnet 4.5 for complex multi-column / table-heavy layouts, Claude Haiku 4.5 for lightweight extraction. The deterministic forms pipeline anchors fields to the PDF's embedded AcroForm widgets, restores section headers from shaded/bold heading bands, merges character-comb fields (driver's-license / SSN digit boxes) into single inputs, synthesizes missing signature fields, and renders source tables as fillable accessible grids. A feedback loop turns each reviewer correction into a safe, per-file deterministic directive (delete / rename a specific field) applied on the next run — replacing the statistical consensus templates that used to force-fit and regress forms. A read-only Pipeline Advisor with persistent memory gathers windowed evidence, surfaces findings, and tracks whether each shipped fix actually moved the metric — without ever mutating prompts or config. 734 automated tests guard the pipeline.
Impact & Results
Key Features
Next Project
File Intelligence Platform