texguardian 0.3.5

Claude Code-style terminal chat app for LaTeX papers

Claude Code, but for LaTeX academic papers

Installation · Quick Start · Commands · Configuration · Guide

---

The Problem

Every conference deadline, the same painful ritual repeats. Your figure overflows the margin at width=1.4\columnwidth. You used bare \cite{} instead of \citep{} — 47 times. Three TODO markers are still in the paper. The bibliography has a hallucinated reference your co-author added from ChatGPT. You forgot to anonymize the acknowledgments section. And it's 2 AM.

You've spent more time fighting LaTeX formatting than writing the actual research.

The Solution

TexGuardian is an interactive CLI that reads your entire paper, understands venue requirements, and fixes every issue through reviewable diff patches — so you never lose work.

One command audits your entire paper:

texguardian chat
>>> /review full

It runs a 7-step pipeline: compile → verify → fix → validate citations → analyze figures → analyze tables → visual polish. Every edit is shown as a unified diff you approve before it touches your files. Automatic checkpoints mean you can always roll back.

Think of it as Claude Code, but for LaTeX.

Highlights

• 26 slash commands covering every stage of paper preparation
• LLM-powered fixes — figures, tables, citations, sections, anonymization, camera-ready
• Real citation validation — checks your .bib against CrossRef and Semantic Scholar, catches hallucinated and outdated references
• Visual polish loop — renders your PDF to images, sends pages to a vision model, catches overlapping figures and spacing issues text analysis misses
• Unified diff patches — every edit is reviewable, with checkpoints and instant rollback
• Natural language — ask for anything in plain English, or use slash commands
• 14 venue templates — NeurIPS, ICML, ICLR, AAAI, CVPR, ACL, ECCV, and more
• Two LLM providers — AWS Bedrock or OpenRouter, switch with one command

Installation

From PyPI (recommended)

pip install texguardian

From source

git clone https://github.com/arcAman07/TexGuardian.git
cd TexGuardian
pip install -e ".[dev]"

External tools

TexGuardian needs LaTeX and Poppler installed on your system for compilation and visual checks:

Tool	Purpose	Install
Python 3.11+	Runtime	python.org
LaTeX (latexmk)	Compilation	See below
Poppler	PDF rendering for visual checks	brew install poppler (macOS) / apt install poppler-utils (Ubuntu)

LaTeX installation

Option A: TinyTeX (~250 MB) — recommended:

# macOS / Linux
curl -sL "https://yihui.org/tinytex/install-bin-unix.sh" | sh

# Add to PATH (add to your ~/.zshrc or ~/.bashrc):
export PATH="$HOME/Library/TinyTeX/bin/universal-darwin:$PATH"  # macOS
export PATH="$HOME/.TinyTeX/bin/x86_64-linux:$PATH"             # Linux

# Install common academic packages:
tlmgr install booktabs natbib hyperref pgfplots xcolor float geometry \
              amsmath amssymb graphicx tikz caption subcaption

TinyTeX is a minimal, portable TeX Live distribution. Missing packages are auto-installed on first compile, or install manually with tlmgr install <pkg>. See yihui.org/tinytex for details.

Option B: Full TeX Live (~4 GB) — includes every package, no missing-package issues:

# macOS
brew install --cask mactex-no-gui

# Ubuntu/Debian
sudo apt install texlive-full

Verify installation

texguardian doctor

TexGuardian Doctor

Checking external tools...

  ✓ latexmk      ~/Library/TinyTeX/bin/universal-darwin/latexmk
  ✓ pdflatex     ~/Library/TinyTeX/bin/universal-darwin/pdflatex
  ✓ pdfinfo      /opt/homebrew/bin/pdfinfo
  ✓ pdftoppm     /opt/homebrew/bin/pdftoppm

All tools found. You're good to go!

Quick Start

1. Initialize your project

cd /path/to/your/latex/paper
texguardian init

This creates three files:

File	Purpose
texguardian.yaml	LLM provider, model, safety limits, LaTeX settings
paper_spec.md	Venue, deadline, thresholds, custom checks, system prompt
.texguardian/	Runtime data (checkpoints, history)

2. Configure credentials

AWS Bedrock (recommended):

# Create a .env file in your project root:
AWS_ACCESS_KEY_ID=your-key
AWS_SECRET_ACCESS_KEY=your-secret
AWS_REGION=us-east-1

OpenRouter:

# In texguardian.yaml:
providers:
  default: "openrouter"
  openrouter:
    api_key: "sk-or-..."  # Get from https://openrouter.ai/keys

3. Start chatting

texguardian chat

On startup you'll see a styled welcome panel with paper stats:

╭──────────────────────────────────────────────────────────╮
│  TexGuardian                                             │
│                                                          │
│  Paper  Scaling Sparse MoE for Long-Context Doc...       │
│  Venue  NeurIPS 2026          Deadline  2026-05-15       │
│  Model  claude opus 4.5       Provider  bedrock          │
│  File   demo_paper.tex        Figures 4 · Tables 3       │
│                                                          │
│  Type /help for commands or ask a question.              │
╰──────────────────────────────────────────────────────────╯

❯ _

Then interact with commands or plain English:

❯ /figures fix
Figure Pipeline

Step 1: Verifying Figures
  Found 4 figures, 3 issues detected

Step 2: LLM Fix
  Analyzing figure issues...

  ✓ Generated 3 patches:
    1. demo_paper.tex (+1/-1) — width=1.3\columnwidth → width=\columnwidth
    2. demo_paper.tex (+1/-1) — width=1.8\textwidth → width=\columnwidth
    3. demo_paper.tex (+1/-1) — width=2.0\columnwidth → width=\columnwidth

  [A]pply all | [R]eview | [N]o: a
  ✓ Checkpoint created
  ✓ All patches applied

❯ The tables use \hline — please fix them to use booktabs

Commands

Paper Analysis

Command	Aliases	Description
/verify	/v, /check	Run all verification checks (figures, citations, custom rules)
/figures [verify|fix]	/figs, /fig	Verify and fix all figures
/tables [verify|fix]	/tabs, /tab	Verify and fix all tables
/citations [verify|fix]	/cite, /refs	Verify, validate (CrossRef/S2), and fix citations
/section <name> [fix]	/sec	Deep analysis of a specific section
/page_count	/pages, /pc	Page count with section breakdown and limit check
/feedback	—	Comprehensive paper review with scoring
/suggest_refs	/suggest_citations	AI-powered citation recommendations

Paper Preparation

Command	Aliases	Description
/anonymize	/anon, /blind	Make paper anonymous for double-blind review
/camera_ready	/cr, /final	Convert draft to camera-ready version
/venue <name> [year]	/template, /conf	Download conference LaTeX templates
/compile	/c, /build	Compile LaTeX document with latexmk
/review	/full, /pipeline	Full pipeline: compile -> verify -> fix -> repeat
/polish_visual	/pv, /visual	Visual verification with vision model

Configuration

Command	Aliases	Description
/model	/m	Show current model configuration
/model list		List all available models
/model set <name>		Change the LLM model
/model search <query>		Search for models by name

File Operations

Command	Aliases	Description
/read <file>	/cat	Display file contents
/write <file>	—	Write content to a file
/grep <pattern>	/g	Search for pattern in files
/search <pattern>	/find, /ls	Find files by name
/bash <cmd>	/sh, /!	Run a shell command

Version Control

Command	Aliases	Description
/diff	/d	Show changes since last checkpoint
/revert	/undo, /rollback	Revert to a previous checkpoint
/approve	/apply, /a	Apply pending patches
/watch	/w	Toggle auto-recompile on file changes

Other

Command	Aliases	Description
/help	/h, /?	Show all available commands
/report	/r	Generate a comprehensive verification report

Natural Language

You can also just type in plain English. TexGuardian will understand:

❯ The figure on line 44 has width=1.3\columnwidth which causes overflow. Please fix it.
❯ Please make this paper anonymous for double-blind review
❯ Can you suggest more citations for the related work section?

The /venue and /model commands also accept natural language:

❯ /venue please download the neurips 2026 style files
❯ /model we need to switch to claude sonnet 4 on bedrock

Configuration

texguardian.yaml

project:
  main_tex: "main.tex"          # Your main .tex file
  output_dir: "build"           # Output directory for PDF

providers:
  default: "bedrock"            # "bedrock" or "openrouter"
  bedrock:
    region: "us-east-1"
    # Credentials loaded from .env automatically
  openrouter:
    api_key: "sk-or-..."
    base_url: "https://openrouter.ai/api/v1"

models:
  default: "claude opus 4.5"    # Main model for analysis and fixes
  vision: "claude opus 4.5"     # Vision model for visual polish

safety:
  max_changed_lines: 50         # Max lines per patch
  max_rounds: 10                # Max auto-fix iterations
  max_visual_rounds: 5          # Max visual verification loops
  allowlist:                    # Files the LLM can modify
    - "*.tex"
    - "*.bib"
    - "*.sty"
    - "*.cls"
  denylist:                     # Files the LLM cannot touch
    - ".git/**"
    - "*.pdf"
    - "build/**"

latex:
  compiler: "latexmk"           # Build tool
  engine: "pdflatex"            # TeX engine
  shell_escape: false           # Enable --shell-escape
  timeout: 240                  # Compilation timeout (seconds)

visual:
  dpi: 150                      # PDF render resolution
  diff_threshold: 5.0           # Convergence threshold (%)
  pixel_threshold: 15           # Per-pixel diff threshold (0-255)
  max_pages_to_analyze: 0       # 0 = all pages

paper_spec.md

The paper spec defines venue-specific rules, custom checks, and your system prompt:

---
title: "Your Paper Title"
venue: "NeurIPS 2026"
deadline: "2026-05-15"
thresholds:
  max_pages: 9
  min_references: 30
  max_self_citation_ratio: 0.2
human_review:
  - "Changes to abstract"
  - "Deletion of more than 10 lines"
---

## Custom Checks

```check
name: citation_format
severity: warning
pattern: "\\cite\\{(?!p|t)"
message: "Use \\citep{} or \\citet{} instead of \\cite{}"
```

```check
name: todo_remaining
severity: error
pattern: "TODO|FIXME|XXX"
message: "Remove TODO/FIXME markers before submission"
```

## System Prompt

```system-prompt
You are an expert academic writing assistant specializing in machine learning.
Use formal academic English. Every claim must be backed by evidence.
```

Environment Variables

Variable	Default	Description
AWS_ACCESS_KEY_ID	—	AWS access key for Bedrock
AWS_SECRET_ACCESS_KEY	—	AWS secret key for Bedrock
AWS_REGION	us-east-1	AWS region
TEXGUARDIAN_MAX_CONTEXT_TOKENS	100000	Max conversation context tokens
TEXGUARDIAN_SUMMARY_THRESHOLD	80000	Token threshold for auto-compaction
TEXGUARDIAN_MAX_OUTPUT_TOKENS	32000	Max LLM output tokens
TEXGUARDIAN_MAX_THINKING_TOKENS	16000	Max thinking/reasoning tokens

Available Models

AWS Bedrock

Friendly Name	Model ID
claude opus 4.5	us.anthropic.claude-opus-4-5-20251101-v1:0
claude opus 4	us.anthropic.claude-opus-4-20250514-v1:0
claude sonnet 4	us.anthropic.claude-sonnet-4-20250514-v1:0
claude-3.7-sonnet	us.anthropic.claude-3-7-sonnet-20250219-v1:0

OpenRouter

Friendly Name	Model ID
claude opus 4.5	anthropic/claude-opus-4.5
claude opus 4	anthropic/claude-opus-4
claude sonnet 4	anthropic/claude-sonnet-4
gpt-4o	openai/gpt-4o

Plus any model from openrouter.ai/models.

Development

# Run tests
pytest tests/ -v

# Run tests with coverage
pytest tests/ --cov=texguardian --cov-report=term-missing

# Lint
ruff check src/

# Type check
mypy src/texguardian

See docs/GUIDE.md for the full architecture and project layout.

License

MIT License. See LICENSE for details.