Metadata-Version: 2.4
Name: paperforge-mcp
Version: 1.0.2
Summary: MCP server: research materials → submission-ready LaTeX papers
Author: Sandi Htun
License-Expression: MIT
Requires-Python: >=3.11
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: fastmcp>=2.0.0
Requires-Dist: anthropic>=0.40.0
Requires-Dist: python-dotenv>=1.0.0
Requires-Dist: httpx>=0.27.0
Requires-Dist: python-pptx>=1.0.0
Requires-Dist: pypdf>=5.0.0
Requires-Dist: pdfplumber>=0.11.0
Requires-Dist: PyGithub>=2.4.0
Requires-Dist: GitPython>=3.1.40
Requires-Dist: bibtexparser>=1.4.0
Requires-Dist: matplotlib>=3.8.0
Provides-Extra: transcription
Requires-Dist: openai-whisper>=20231117; extra == "transcription"
Provides-Extra: overleaf
Requires-Dist: pyoverleaf>=0.1.0; extra == "overleaf"
Provides-Extra: cloud
Requires-Dist: fastapi>=0.115.0; extra == "cloud"
Requires-Dist: uvicorn[standard]>=0.30.0; extra == "cloud"
Requires-Dist: pydantic>=2.0.0; extra == "cloud"
Provides-Extra: dev
Requires-Dist: pytest>=8.0.0; extra == "dev"
Requires-Dist: pytest-asyncio>=0.24.0; extra == "dev"
Requires-Dist: fastapi>=0.115.0; extra == "dev"
Requires-Dist: uvicorn[standard]>=0.30.0; extra == "dev"
Requires-Dist: pydantic>=2.0.0; extra == "dev"
Dynamic: license-file

# PaperForge MCP

An MCP server that turns research materials into venue-styled LaTeX papers. Feed it your slides, PDFs, code repos, notes, and figures -- it produces submission-ready drafts for IEEE, NeurIPS, ACM, ICML, Springer, and more.

## Install

```bash
pip install paperforge-mcp
```

Requires Python 3.11+.

## Quick start

Add to your MCP client config (Claude Desktop, Cursor, Windsurf, etc.):

```json
{
  "mcpServers": {
    "paperforge": {
      "command": "paperforge-mcp"
    }
  }
}
```

Then ask your AI assistant to write a paper -- PaperForge handles the rest.

## What it does

**Research pipeline:**
- Ingest PDFs, PowerPoint slides, markdown notes, images, audio/video, and GitHub repos
- Fetch LaTeX templates for 9+ venues (IEEE, NeurIPS, ACM SIGCONF, ICML, ICLR, Springer LNCS, Elsevier)
- Build citation libraries from CrossRef and Semantic Scholar
- Extract and validate algorithms and equations from source materials

**Paper generation:**
- Generate structured outlines before full drafts
- Produce complete LaTeX papers with proper venue formatting
- Auto-place figures, tables, and citations
- Validate LaTeX syntax, structure, and content quality
- Link every claim back to source evidence for traceability

**Publishing:**
- Export ready-to-submit ZIP packages
- Publish directly to Overleaf

## Tools

| Tool | Description |
|------|-------------|
| `ingest_research_materials` | Scan and extract content from PDFs, slides, notes, images, video, GitHub repos |
| `read_github_repository` | Clone and analyze a GitHub repository |
| `fetch_journal_template` | Download LaTeX template for your target venue |
| `build_citation_library` | Fetch references from CrossRef / Semantic Scholar |
| `configure_paper` | Set title, authors, venue, and preferences |
| `preview_outline` | Generate a structured outline before writing |
| `generate_paper_draft` | Create the full paper with all sections |
| `assemble_draft` | Compile sections into a complete document |
| `validate_latex` | Review and auto-repair LaTeX quality |
| `extract_and_validate_algorithms` | Extract algorithms and equations from sources |
| `extract_equations_from_slides` | Parse math equations from PPTX slides |
| `generate_table` | Create LaTeX tables from CSV data or descriptions |
| `link_claims_to_evidence` | Map paper claims to source materials |
| `refine_section` | Apply feedback to specific sections |
| `publish_to_overleaf` | Package and upload to Overleaf |

## Cloud tier

For server-side LLM generation, LaTeX compilation, AI-powered repair, and smart citation suggestions, see the cloud tier at [paperforge.dev](https://paperforge-production-af9f.up.railway.app/docs).

## License

MIT
