DocuForge
High-fidelity document intelligence: convert legacy ODF and technical PDFs into structured JSON or Markdown with free BYOK.
Millions of critical enterprise documents, technical blueprints, and public administration archives remain locked in legacy Open Document Format files (ODF: .odt text documents, .ods spreadsheets, .odp presentations) and complex multi-column PDFs. Ingesting these archives into modern LLM pipelines, knowledge graphs, or vector databases frequently results in broken tabular layouts, scrambled cross-references, or prohibitive SaaS subscription costs.
DocuForge solves this by providing a private, high-precision document extraction engine with a completely free BYOK (bring your own key) model. You connect your own preferred AI model API key (such as OpenAI, Anthropic, Google Gemini, or Mistral) or point to a local self-hosted LLM (like Ollama or vLLM) on your private network. You pay zero software subscription fees to DocuForge, paying only your direct token cost or running completely free on local hardware.
DocuForge reconstructs document structure deterministically: complex multi-level tables are converted into clean semantic tables or typed JSON arrays, nested section hierarchies are preserved, and technical formulas and metadata are extracted with full audit traceability back to the original document page coordinates. Whether modernising legacy public sector archives or preparing data for enterprise RAG, DocuForge delivers reliable, machine-readable outputs with zero vendor lock-in.
Key Features
Architectural Highlights
Free BYOK architecture
Use the application completely free with your own AI API keys or local on-premise models, with zero software licensing or per-page subscription fees.
Comprehensive ODF and PDF conversion
Seamlessly parse legacy Open Document Format files (.odt, .ods, .odp) and complex scanned technical PDFs into machine-readable structures.
Structured JSON and clean Markdown
Generate strictly validated JSON schemas or clean GitHub-flavoured Markdown ready for vector databases, LLM ingestion, or archival databases.
Audit-ready coordinate traceability
Every extracted value, table cell, and paragraph is mapped back to its visual coordinate and source page for full audit compliance.
Engineering Specs
Technical Profile
Technology Stack
Support & Details
Frequently Asked Questions
What document formats does DocuForge support? +
DocuForge natively supports Open Document Format files (.odt, .ods, .odp), complex technical PDFs, scanned high-resolution TIFF/PNG drawings, and legacy DOCX files.
How does the free BYOK (bring your own key) model work? +
DocuForge is free to use without software licenses or page limits. You simply input your own API key for OpenAI, Anthropic, Google Gemini, or Mistral, or point to a free local LLM running in Ollama or vLLM. You pay only your direct API provider token costs, or run completely free on local hardware.
Does DocuForge store or train on my uploaded documents? +
No. DocuForge operates with an in-memory, zero-retention pipeline. Documents are processed in volatile memory and purged immediately upon conversion. We never store your documents or use them for model training.
Can I run DocuForge in an air-gapped on-premise environment? +
Yes. DocuForge is available as a self-hosted Docker container that runs completely offline with local models (e.g. Ollama or vLLM), ensuring strict compliance for defense, legal, and public sector deployments.
Official Documentation & Support
Support contact: enterprise@4dc.services