LexVaultMD — Legal PDF to Markdown CLI for Litigation Teams

LexVaultMD by Nexacrawl converts legal PDFs into clean, structured, AI-ready Markdown for discovery, exhibits, pleadings, case files, and litigation review.

Repository: github.com/Kevork-Nexacrawl-dev/lex-vault-md

Why LexVaultMD?

Legal PDFs — motions, briefs, exhibits, deposition transcripts, discovery productions — are structurally rich but extraction-hostile. LexVaultMD reads the PDF transform matrix to detect real heading hierarchy, strips headers/footers/watermarks, detects multi-column layouts, extracts tables, and runs entirely offline with zero API calls. The output is clean, structured Markdown that is immediately usable for AI-assisted review, case management, and litigation workflows.

| | LexVaultMD (Nexacrawl) | Typical converters | |---|---|---| | Speed | ~2 seconds for a 12 MB PDF | Often slower; many use cloud APIs | | Headings | Font-size-based detection from transform matrix | Flat text or guesswork | | Modes | local, web, batch | Usually single-file only | | Offline | Yes — no internet required for local/batch | Varies | | Output | Structured .md with page markers, metadata header | Plain text dump | | Legal Artifacts | Header/footer removal, watermark stripping, margin noise | Not handled |

Install

npm install -g @nexacrawl/lex-vault-md

Requires Node.js 18+.

Quick Start

# Convert a local legal PDF
lex-vault-md local ./motion.pdf

# Fetch and convert a remote PDF
lex-vault-md web https://example.com/brief.pdf

# Batch convert an entire discovery folder
lex-vault-md batch ./discovery/ --output ./case-vault/

Usage

Convert a local PDF

lex-vault-md local ./motion.pdf

Fetch and convert a remote PDF

lex-vault-md web https://example.com/brief.pdf

Batch convert an entire folder

# Convert all PDFs — output .md files to the same folder
lex-vault-md batch ./discovery/

# Write converted files to a separate case vault folder
lex-vault-md batch ./discovery/ --output ./case-vault/

# Control how many PDFs process in parallel (default: 3)
lex-vault-md batch ./discovery/ --output ./case-vault/ --concurrency 5

Batch output example:

ℹ Found 10 PDF(s) in discovery
ℹ Output → ./case-vault

  ✔ exhibit-001.pdf → exhibit-001.md  (4 page(s), 7203 chars)
  ✔ motion-to-dismiss.pdf → motion-to-dismiss.md  (3 page(s), 5891 chars)
  ✔ deposition-smith.pdf → deposition-smith.md  (2 page(s), 3142 chars)
  ⏭ exhibit-002.pdf — Skipped (exists)
  ✖ sealed-filing.pdf — encrypted PDF — skipped

────────────────────────────────────────────────
  ✔ Batch complete — 10 file(s) processed
  ✔ Converted : 8
  ⏭ Skipped   : 1  (already exists)
  ✖ Failed    : 1

Batch mode is safe to re-run — already-converted files are skipped automatically.

Flags

| Command | Flag | Short | Description | |---|---|---|---| | local | --output <file> | -o | Custom output filename (default: same name as PDF) | | local | --clipboard | -c | Copy Markdown to clipboard after saving | | web | --output <file> | -o | Custom output filename | | web | --clipboard | -c | Copy Markdown to clipboard after saving | | batch | --output <dir> | -o | Output directory for .md files (default: same as source folder) | | batch | --concurrency <n> | -n | Number of PDFs to process in parallel (default: 3) |

Examples

# Single file — save to a custom path
lex-vault-md local ./pleading.pdf --output ./vault/pleading.md

# Fetch remote PDF and copy result to clipboard
lex-vault-md web https://example.com/exhibit.pdf --clipboard

# Batch with higher concurrency for large discovery productions
lex-vault-md batch ./production/ --output ./review/ --concurrency 8

# Full example with both flags
lex-vault-md local ./filing.pdf -o ./filings/filing.md -c

Performance

LexVaultMD is built for speed — pure Node.js, local extraction, no cloud round-trips.

| Document | Result | |---|---| | 12 MB PDF | ~2 seconds on a typical machine | | Batch folders | Parallel processing with --concurrency (default: 3) | | Network | Not required for local and batch modes |

Output Format

Every conversion produces a structured Markdown file:

<!-- Source: motion.pdf | Extracted: Jun 29, 2026, 10:00 AM -->

[page 1 content flows here]

<!-- page 2 -->
*— page 2 —*

# ARGUMENT

## I. THE COURT HAS JURISDICTION

Paragraph text flows here with proper line breaks
and paragraph spacing preserved from the original PDF.

Heading detection maps font sizes from the PDF's transform matrix:

| Font size | Markdown heading | |---|---| | ≥ 22px | # H1 | | ≥ 18px | ## H2 | | ≥ 15px | ### H3 | | > 14px (distinct size) | #### H4 |

Features

• Lightning-fast conversion — local PDFs process in seconds, not minutes
• Smart heading detection — reads raw font-size data from the PDF transform matrix
• Multi-column layout detection — MER algorithm + XY-Cut++ fallback for complex legal layouts
• Table extraction — hybrid lattice/stream orchestrator for bordered and borderless tables
• Artifact removal — strips repeating headers/footers, watermarks (DRAFT, CONFIDENTIAL, COPY), and margin noise
• Garbage gate — detects and skips scanned/image-only pages automatically
• Y-sorted text extraction — items sorted by visual position (top→bottom, left→right)
• Subtle page markers — invisible HTML comment + *— page N —* italic; no large headings
• Source metadata — output header includes filename/URL and extraction timestamp
• Batch conversion — convert entire folders with parallel processing and a clean summary
• Skip on exist — batch mode skips already-converted files; safe to re-run anytime
• Colored terminal output — chalk-powered ✔ success, ✖ error, ⚠ warn messages
• Spinner feedback — ora spinner shows live progress per file
• Graceful error handling — specific messages for missing files, encrypted PDFs, 404s, timeouts
• Works offline — local and batch modes require no internet connection

Error Handling

| Scenario | Behavior | |---|---| | File not found | Clear error with resolved path | | Not a .pdf file | Rejects before attempting extraction | | Encrypted / password-protected PDF | Specific error message | | Scanned / image-only PDF | Detects via garbage gate, skips with placeholder | | Unreachable URL | Distinguishes DNS failure vs. refused connection | | HTTP 404 / 403 | Status-specific error messages | | Request timeout (30s) | Clean timeout message | | Empty folder (batch) | Warns and exits cleanly | | Mixed folder (batch) | Skips non-PDF files silently |

Tech Stack

| Package | Version | Role | |---|---|---| | commander | ^12 | CLI commands & flags | | pdf-parse | ^1.1 | PDF text extraction (Node.js) | | axios | ^1.7 | Fetch remote PDFs | | chalk | ^5 | Colored terminal output | | ora | ^8 | Loading spinner |

Migration from pdf2md-cli

This project was formerly published as pdf2md-cli.

New install:

npm install -g @nexacrawl/lex-vault-md

New CLI command:

lex-vault-md

The legacy pdf2md command remains available temporarily for backward compatibility.

Local Development

git clone https://github.com/Kevork-Nexacrawl-dev/lex-vault-md.git
cd lex-vault-md
npm install
npm link          # makes both lex-vault-md and pdf2md available globally

lex-vault-md local ./test.pdf
lex-vault-md batch ./test-folder/

FAQ

Where is the official repo? → https://github.com/Kevork-Nexacrawl-dev/lex-vault-md

Is this the same as opengovsg/pdf2md? → No. That is a different project. Search for Kevork-Nexacrawl-dev/lex-vault-md or lex-vault-md nexacrawl.

What happened to pdf2md-cli? → This project was renamed to LexVaultMD (@nexacrawl/lex-vault-md) and repositioned for legal teams. The pdf2md CLI alias still works temporarily for backward compatibility.

Who maintains this? → Kevork (Nexacrawl) — contact: [email protected]

License

Business Source License 1.1 — © 2026 Nexacrawl

This project is licensed under the BSL 1.1, which allows non-production use, modification, and redistribution. Production/commercial use requires a separate license.

• Change Date: 2029-01-01 — after this date, the code automatically converts to MIT
• For commercial licensing inquiries: [email protected]

See the LICENSE file for full terms.

Contributing

PRs are welcome. Open an issue first for any significant changes.

Please review CONTRIBUTING.md before opening a pull request, and follow our Code of Conduct in all project spaces.