TeXRA: AI TeX Research Assistant for VS Code

TeXRA brings large language models to rigorous scientific workflows. The extension embeds research-grade agents directly in VS Code so you can draft, review, and manage LaTeX projects without leaving your editor.

See texra.ai or the full documentation for tutorials and a web-based launch page.

Why TeXRA

Reliable scientific workflow – orchestrate literature review, drafting, and revision with reproducible agent runs, structured logs, and built-in verification tools like latexdiff and texcount.
Specialized research agents – deploy focused agents for correcting LaTeX, polishing prose, generating figures, and document transformation—without losing control over references and context.
Transparent reasoning loops – inspect live reasoning, replay sessions, and compare outputs across models to build trust in the generated work.
Model flexibility with guardrails – connect to OpenAI, Anthropic, DeepSeek, Gemini, Moonshot, Qwen, and custom endpoints while keeping API routing and cost monitoring under your control.

Quick Start

Install the extension from the VS Code Marketplace.
Run TeXRA: Create Sample Project from the command palette to explore a fully configured workspace.
Open the TeXRA sidebar, pick an agent, select your files, and click Execute. Follow-up chat is available for tool-use agents.

Requirements

LaTeX distribution (TeX Live, MiKTeX, or MacTeX)
Perl (required for LaTeX formatting tools)
Optional: GraphicsMagick/ImageMagick and Ghostscript for PDF and image processing

Installation

Set your API keys in VS Code settings or create a workspace .env file:

OPENAI_API_KEY=your_openai_key_here
ANTHROPIC_API_KEY=your_anthropic_key_here
GEMINI_API_KEY=your_gemini_key_here

TeXRA loads the .env file automatically at startup. For detailed setup instructions, see the installation guide.

Customization

Configure available agents, prompts, and model parameters in VS Code settings. You can tailor directories, file types, output locations, and more to suit your workflow. Advanced users can define custom agents in YAML or extend the extension with new model handlers.

Support & Feedback

Report issues and feature requests on the GitHub issues page or email contact@texra.ai.

TeXRA

TeXRA.ai