Summary

Paper Burner X is an open-source, browser-based AI workstation designed to streamline research workflows by tackling complex document processing, translation, and analysis. It targets postgraduate students and researchers, offering a unified, privacy-focused environment to overcome barriers presented by extensive PDF literature, intricate formulas, and multilingual content, thereby enhancing efficiency and insight extraction.

How It Works

The project employs a novel frontend Agentic RAG (Retrieval Augmented Generation) system, empowering AI with a global understanding of document structure and a suite of tools (grep, vector search, fetch) for multi-step reasoning and complex information extraction directly within the browser. It features high-performance batch processing for OCR and translation, supporting extensive term banks for consistency. The architecture prioritizes extensibility, allowing integration with custom AI model endpoints and future offline deployment options.

Quick Start & Requirements

• Mode 1: Pure Frontend Deployment (Recommended for personal use)
  • Install: Fork the repository and deploy to Vercel (approx. 5 minutes).
  • Prerequisites: API keys for OCR services (e.g., MinerU, Doc2X) and translation models (e.g., DeepSeek, Gemini, Claude, custom endpoints). Data is stored locally in the browser (localStorage/IndexedDB).
  • Links: Vercel Deployment Guide, Online Demo.
• Mode 2: Docker Full Deployment
  • Status: Backend development is in progress; this deployment option is not yet available.
  • Prerequisites: Docker (when available).

Highlighted Details

• Speed: Concurrent processing enables rapid translation of long documents, often in seconds.
• Format Preservation: Maintains original formatting, including complex formulas, charts, and tables, during OCR and translation.
• AI-Powered Analysis: Features an AI chat assistant for document Q&A, mind map generation, and flowchart creation.
• Privacy & Security: Pure frontend mode ensures all data remains localized within the user's browser.
• Broad Format Support: Imports PDF, DOCX, PPTX, EPUB, Markdown, and code comments; exports to DOCX, MD, HTML, PDF.

Maintenance & Community

The project is under active iteration, with a roadmap including Docker deployment, multi-user systems, and UI refactoring. Contribution is welcomed via bug reports, feature suggestions, and pull requests. Links to community channels like Discord or Slack are not explicitly provided, but the GitHub repository serves as the primary hub.

Licensing & Compatibility

The project is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0). This strong copyleft license requires that any modifications or derivative works, when deployed as a network service (SaaS, internal enterprise service), must also be made available under the AGPL-3.0, including providing users with access to the source code.

Limitations & Caveats

The Docker deployment option is currently under development. AI translation results are for reference only, and OCR accuracy for complex PDF formats may necessitate manual review. Clearing browser data in the pure frontend mode will result in the loss of local history and configurations. Users must adhere to the terms of service for any third-party APIs used.