AutoIllustrate: Automated Generation of Polished and Editable Paper Figures | Tokyo .
Cities (231) ▾

Members-Only

Recent Talks & Demos are for members only

Exclusive feed

You must be an AI Tinkerers active member to view these talks and demos.

Sign in View FAQ
June 18, 2026 · Tokyo

AutoIllustrate: Editable Academic Figures

Learn how to generate publication-grade academic figures from text. This talk details a two-step process for creating editable `.drawio` files and polished PNGs, addressing common issues with direct diffusion model generation.

Claude Code CLI Draw OpenAI API Google Gemini API MathJax
Overview

AutoIllustrate is a command-line tool that turns a methodology brief (a paper section, a system description, or a verbal sketch) into a publication-grade academic figure of 1 to 3 panels. Each run produces two deliverables: (1) an editable .drawio file and (2) an enhanced, ready-to-publish image.

Live, I’ll walk through the full pipeline run on a real paper section, stage by stage:

  • the agent deciding the layout and panel count,
  • each panel rendered as a draw.io diagram,
  • the vision-model audit-and-repair loop,
  • deterministic composition of the panels into a single file,
  • the optional image-model restyle into a camera-ready PNG, followed by the verification and repair steps.
Tech stack
  • Claude Code CLI
    An agentic coding assistant that integrates Anthropic's AI models directly into your terminal for conversational development.
    Claude Code CLI is Anthropic's command-line interface (CLI) for agentic coding: it meets developers where they work, in the terminal. This tool allows you to build features from plain English descriptions, debug complex issues, and navigate any codebase with conversational commands. It is powerful because it takes action: the CLI can directly edit files, run shell commands, and even manage Git operations like creating commits and branches. Installation is straightforward via `npm install -g @anthropic-ai/claude-code`, and it's designed to be a flexible, scriptable power tool that automates tedious tasks like fixing linting errors or resolving merge conflicts.
  • Draw
    A free, open-source diagramming application built for secure, collaborative technical drawing.
    Draw (widely known as draw.io) is a security-first diagramming tool designed for teams mapping out complex architectures, flowcharts, and network designs. Operating under an Apache 2.0 license, it bypasses the typical enterprise paywalls and artificial scarcity of VC-backed competitors, giving over 100 million users worldwide free access to professional-grade tools. The platform prioritizes absolute data privacy by storing files directly on your preferred cloud storage (such as Google Drive, OneDrive, or GitHub) or locally on your desktop. With native integrations for Atlassian Confluence, Jira, and VS Code, it fits seamlessly into developer workflows, offering robust shape libraries and real-time cursor collaboration without requiring a sign-up.
  • OpenAI API
    OpenAI API: Your direct gateway to cutting-edge AI models (GPT-4o, DALL-E 3, Whisper), enabling scalable, multimodal intelligence integration into any application.
    The OpenAI API provides authenticated, programmatic access to a powerful suite of generative AI models. Developers leverage REST endpoints and official libraries (Python, Node.js) to integrate capabilities like advanced text generation (GPT-4o), image creation (DALL-E 3), and speech-to-text transcription (Whisper). This platform is engineered for scale, supporting millions of daily requests for tasks from complex reasoning to real-time customer support agents, ensuring your application gets reliable, state-of-the-art intelligence.
  • Google Gemini API
    Integrate Google's multimodal Gemini models (Pro, Flash) into your application via REST or SDKs: generate content, process up to 1000-page PDFs, and execute code with a 2-million token context.
    The Gemini API delivers Google's most advanced models (Gemini 3 Pro, 2.5 Flash) directly into your applications. Leverage its multimodal power: process text, images, video, and audio inputs for tasks like content generation, summarization, and visual understanding. Utilize key features including the 2-million token context window, structured JSON output, and function calling to build complex, reliable agents. Choose your integration: use the standard REST API, streaming endpoints (SSE), or the Live API (WebSockets) for real-time conversational experiences. Get started with the free tier and robust SDKs for Python, Java, Go, and more.
  • MathJax
    MathJax is an open-source JavaScript display engine that renders high-quality LaTeX, MathML, and AsciiMath notation directly in any modern web browser.
    MathJax eliminates the need for clunky mathematical images or specialized plugins by converting raw LaTeX, MathML, and AsciiMath markup into crisp, accessible vector graphics (SVG) or HTML-with-CSS typography. Developed as a joint venture under the American Mathematical Society and the Society for Industrial and Applied Mathematics, this highly modular library integrates seamlessly into web platforms to ensure equations scale dynamically with surrounding text. Its built-in support for screen readers and copy-paste functionality makes complex mathematics universally readable and interactive across all devices.
Events Gallery FAQ Sponsor