**AI-Powered Document Automation System — Full Build** I need an experienced developer to build an end-to-end document generation system powered by AI. I upload a package of files, the system reads and extracts everything automatically, AI writes all the copy in my voice, I make any quick edits in a live preview, and a print-ready branded PDF comes out the other side. Fast. No manual data entry. **How it works:** 1. I upload a package of documents and photos 2. I select the asset type from a simple dropdown 3. System automatically reads and extracts all relevant data 4. AI writes all narrative copy in my voice — tone and messaging automatically adjusts based on the asset type selected 5. Live preview generated I click in, make any quick edits to text, data, or photos 6. I approve it, PDF downloads **What you'll build:** - Simple upload interface with asset type selector - Document parsing and data extraction including scanned files and OCR - AI integration with asset-type-aware copy generation - Live editable preview before PDF locks — this is a required feature, not optional - Pixel-accurate branded PDF output — fonts, colors, logo placement, layouts locked to provided brand standards **What I'll provide:** - Corporate branding cheat sheet hex colors, fonts, spacing, logo usage rules - Official logo files in all required formats - Full design reference PDFs as layout guides - Copy style guide and voice examples for AI calibration - Sample input files showing exactly what I upload each time --- **Who I'm looking for:** - Strong experience with LLM APIs — Anthropic/Claude preferred, open to best approach - OCR and document parsing experience — AWS Textract or Google Document AI preferred - Proven PDF generation experience — Puppeteer, WeasyPrint, ReportLab, or equivalent - Has built a live editable preview layer before, show me an example - Python automation background preferred - This is a replication and automation project, not a creative one - Propose the right tool for each component, not just what you know best --- **To apply:** 1. Show me examples of AI-powered document automation you've built 2. Show me examples of design-accurate branded PDF output 3. Tell me specifically how you'd approach the live editable preview 4. Give me your recommended tech stack by component 5. Fixed-price quote with milestone breakdown No examples, no review.