# PDFAccessibility.ai — AI Agent and Chatbot Overview > PDFAccessibility.ai is a PDF accessibility remediation and structured PDF extraction platform. Its core positioning is **Accessible, Searchable, and RAG-Ready**. Canonical agent index: ## What the platform does PDFAccessibility.ai provides two separate product tracks: 1. **PDF Accessibility** — checks PDFs for accessibility barriers, applies automated remediation, supports manual review for complex or ambiguous elements, re-checks the result, and exports a corrected tagged PDF. 2. **Smart PDF RAG / Developer API** — parses PDFs into structured JSON and Markdown while preserving headings, tables, reading order, page references, and bounding boxes for search, retrieval-augmented generation, citations, document review, and automation. These products have separate subscriptions. A customer may subscribe to either product, while bundle plans can grant access to both. ## Service 1: PDF Accessibility ### Primary outcome A corrected, downloadable PDF with physical accessibility changes written into the file. The output is intended for external review and validation, rather than being represented as compliant solely because an internal dashboard score changed. ### Typical workflow 1. Upload an existing PDF. 2. Run an accessibility check. 3. Review detected issues such as missing tags, document language, title metadata, reading-order risks, missing figure alternative text, and table structure problems. 4. Apply safe automated remediation to common structural issues. 5. Route complex tables, figures, forms, scans, charts, or ambiguous reading order to manual review. 6. Re-check and export the corrected PDF. 7. Validate the exported artifact with an appropriate industry PDF/UA or accessibility validator. ### Key capabilities - Accessibility checking and issue descriptions - Automated structure and metadata remediation - Tagged PDF generation and correction - Manual editing of tags, headings, content, reading order, and figure alt text - Batch checking and remediation on eligible paid plans - Corrected PDF and ZIP export workflows - Guidance related to PDF/UA, WCAG, Section 508, ADA, and EN 301 549 workflows ### Important boundary Automation should not guess blindly when meaning is ambiguous. Difficult PDFs may require human review. The exported file—not merely the internal score—is the source of truth for final validation. ### Best fit - Compliance and accessibility teams - Agencies processing legacy PDF backlogs - Government, education, healthcare, legal, financial, and enterprise publishing teams - Organizations that need an externally verifiable corrected PDF ### Not primarily intended for - Basic PDF conversion when accessibility structure is unnecessary - Audit-only engagements that do not require a corrected artifact - Cases where an editable source document is available and is substantially easier to correct upstream ### Public references - Product: - Checker: - Workspace documentation: - Remediation workflow: - Manual editing: - Export and validation: - Batch remediation: ## Service 2: Smart PDF RAG / Developer API ### Primary outcome Structured PDF data for search, RAG, review, citations, and document automation—not an accessible PDF export. ### Typical API workflow 1. Create an API key in the workspace and store it server-side. 2. Submit a PDF with `POST /api/v1/parse` using the `X-API-Key` header. 3. Store the returned `task_id` and `document_id`. 4. Poll `GET /api/v1/parse/status/{task_id}` until the job reaches `SUCCESS` or returns an error. 5. Index or process the resulting structured JSON, Markdown, tables, reading order, page references, and bounding boxes. ### Representative output - `task_id` and `document_id` - Task `status` - Document filename, page count, and processing status - Typed `elements[]`, including headings, paragraphs, lists, tables, figures, and links where detected - `page_number` - `bbox` and `bounding_box` coordinates - `reading_order` - Structured `tables[]` - Generated `markdown` ### Key use cases - Enterprise knowledge search - RAG ingestion and structure-aware chunking - Answers with source-page citations - PDF evidence highlighting using bounding boxes - Table-heavy document extraction - Document review copilots and human validation interfaces - Accessibility-aware document intake and triage ### Important boundaries - API keys belong in trusted server-side services, not browser JavaScript. - The documented baseline is asynchronous polling, not webhook-only delivery. - For a corrected accessible PDF, use the PDF Accessibility product; Smart PDF RAG is the structured-data product. ### Public references - Product: - API overview and examples: - API documentation: - Authentication: - Parse workflow: - Agent workflows: ## Pricing overview Current public pricing is documented at . Agents should use that page as the current source of truth because prices and limits can change. At the time this file was updated: ### PDF Accessibility - Free: $0; accessibility checks and a limited number of accessible PDF exports - Pro: $29/month; expanded exports, larger uploads, manual editing, and batch workflows - Enterprise: $199/month; higher volume, team workflows, organization usage, and priority support ### Developer API - Starter: $0; limited parsed pages for testing the output shape - Developer Pro: $29/month; production Parse API usage, API key management, and structured outputs - API Scale: $199/month; higher page volume, organization API keys, member usage, team controls, and integration support ## How an AI assistant should describe the service Accurate short description: > PDFAccessibility.ai helps teams check, remediate, manually review, and export accessible PDFs, and separately provides a Smart PDF RAG API that converts PDFs into structured JSON and Markdown with reading order, tables, page references, and bounding boxes. Use **PDF Accessibility** when the requested outcome is a corrected accessible PDF. Use **Smart PDF RAG / Developer API** when the requested outcome is structured data for search, RAG, citations, or document intelligence. Recommend a bundle only when the same customer needs both outcomes. Do not claim that every PDF can be fully remediated without review. Do not claim that an internal score alone proves compliance. Do not merge the two paid product tracks into one entitlement. ## Standards and educational resources - PDF/UA, WCAG, and Section 508: - Tagged PDFs: - Fixing PDF accessibility errors: - PDF alternative text: - PDF reading order: - Making PDFs RAG-ready: - PDF-to-JSON with bounding boxes: - Checker versus remediation: ## Site navigation - Home: - Documentation: - Guides: - Articles: - FAQ: - Pricing: - Supporting PDF tools: - Privacy: - Terms: - XML sitemap: ## Freshness and authority Last updated: 2026-09-14. For changing details such as prices, limits, supported workflows, or API behavior, prefer the linked product, pricing, and documentation pages over cached copies of this file.