Searchable OCR
OCRmyPDF and Tesseract turn PDFs, JPGs, PNGs, TIFFs, and WebP images into searchable text. Existing text stays intact; scanned pages become useful.
Everything has a place
A complete toolkit for your personal paper trail.
DocManFu helps you capture, understand, organize, retrieve, and protect the documents that quietly run your life.
AI document analysis
Automatically classify bills, statements, medical records, tax forms, insurance documents, and more. Generate sensible names, tags, and structured metadata.
Fast full-text search
Search names and extracted document content, then narrow results by type, tag, upload date, or date range. The phrase you remember is enough.
Tags and smart metadata
Add your own color-coded tags or accept AI suggestions. Edit names and metadata at any time without changing the original file.
Bill tracking
Keep amount, due date, paid status, and payment details beside the bill itself. Quick filters reveal what is upcoming or overdue.
Flexible capture
Drag and drop multiple files, upload from any browser, import Evernote ENEX archives, or automate a scanner folder with the DocManFu CLI watcher.
Built-in PDF preview
Preview documents in the browser, inspect extracted text and AI metadata, edit tags, download the original, or reprocess with updated settings.
Background processing
Celery and Redis process OCR, AI analysis, and imports outside the request cycle. Live job progress and status events keep the interface responsive.
Self-hosted storage
Files remain on storage you select. Deploy to a NAS, homelab, or VPS with Docker and standard open-source infrastructure.
Choose your AI
Use OpenAI, Anthropic, local Ollama, or no AI at all. OCR and document management continue to work when AI is disabled.
Accounts and roles
JWT-based authentication, admin and user roles, per-user document ownership, API keys, and first-run setup protect multi-user installations.
Backups and portability
Standard PostgreSQL backups plus ordinary files on disk mean your archive is understandable, movable, and recoverable without a proprietary cloud export.
The processing pipeline
Every upload gets more useful.
The original file is safely stored first. Then DocManFu creates a tracked processing job, runs OCR, extracts text, and—if enabled—asks your chosen AI provider for classification and metadata.
- Original file is preserved
- Retries handle temporary failures
- Every job has visible status and history
- Reprocess documents when settings improve
1CapturePDF or image
→2OCRSearchable text
→3AnalyzeName, type, tags
→4FindInstant retrieval
Your documents. Your rules.
Ready to master the paper pile?
One Docker Compose command puts a private, searchable document vault on hardware you control.