document-to-latex
CommunityVerification-first conversion of scientific papers (PDF) to clean, compilable LaTeX (.tex). Papers, theses, technical reports — any field. Runs multiple extractors at every step (PyMuPDF, pdfplumber, pdfminer.six, OCR), three vision APIs for figures, and closes the loop with round-trip compile-and-diff plus numeric preservation. The final human-review HTML is curated by the LLM, not by heuristics.
Works With
Quality and Utility come from canonical skill metadata. Community ratings appear separately as stars when subscriber reviews exist.
Subscribe to download
Public previews are available, but downloads, CLI sync, and premium content require an active subscription.
Sign In to Subscribe- Full skill content with deterministic versioning
- CLI sync + local search index
- Access to Jeffrey’s premium library
Save to your account
Saved skills live in your cloud list. Run jsm sync on each machine to pull them locally.
What's New with this Skill
1Updates, enhancements, and release notes for this skill.
New community skill: document-to-latex
Convert scientific PDFs into clean, compilable LaTeX with verification gates at every step. Four extractors run in parallel, three vision APIs handle figures, and a round-trip compile-and-diff catches what every method missed.
- Four parallel extractors with per-page pairwise agreement scoring; flagged pages get LLM-curated review rather than a fixed heuristic threshold
- Three vision APIs cross-check figure descriptions instead of trusting any single model
- Numeric preservation check verifies that values in the source survive the round-trip into LaTeX
- Round-trip compile-and-diff confirms the emitted .tex actually builds before signoff
- Operator-facing review HTML curated by the LLM in step 7b, not by hardcoded scoring rules
This skill walks the full pipeline for turning a scientific PDF into clean LaTeX: ingest the PDF, run every available extractor (PyMuPDF, pdfplumber, pdfminer.six, OCR), compute pairwise agreement per page, and surface the divergent pages for LLM-curated review. Three vision APIs handle figure descriptions. A numeric preservation check catches dropped values. The output is a compilable .tex draft plus an HTML review page that the operator can scan in minutes rather than hours. It runs every extractor that is available, and if missing dependencies leave only one method it says so before going o