Where is the boundary between AutoFigure-Edit and Mermaid?

[Mermaid](https://mermaid.js.org) is a declarative diagramming language for engineering docs, great for maintaining stable diagrams as code inside Markdown and rendering in CI. AutoFigure-Edit is for generating and refining scientific figures directly from long-form natural language text, producing editable SVG object layers and automating layout and consistency polishing while preserving room for manual finishing. For simple flowcharts and lightweight diagrams, Mermaid fits better; for method figures and submission-grade illustrations that require iterative refinement, AutoFigure-Edit fits better.

Does it output raster images or vectors, and can I edit without quality loss?

It centers on editable SVG output, where text, shapes, and arrows remain vector objects rather than pixel layers. That means you can keep refining without re-drawing and without accumulating blur or aliasing artifacts, which is crucial for publication workflows.

Can I match the visual style of a target venue or a reference paper?

A core goal is separating structure from style so styling can be standardized after the structure is stable. In practice, you can distill team conventions such as color palettes, fonts, stroke widths, and spacing into reusable templates so multiple figures share a consistent visual language.

AutoFigure-Edit: Editable SVG Figures, Mermaid Alternative

Pain Points vs Innovation

✕Traditional Pain Points	✓Innovative Solutions
Traditional figure creation depends on manual layout. A single change often forces global realignment, and maintenance cost grows linearly with iterations.	AutoFigure-Edit standardizes the output as losslessly editable SVG object layers so text, arrows, and shapes become traceable, reusable, refinable engineering assets.
Many text-to-figure tools only output rasters or semi-structured artifacts. Editing later introduces blur and misalignment, making publication-grade readability hard.	It separates generation from polishing via iterative layout and style refinement: lock structure first, then optimize aesthetics and consistency, reducing structural drift common in one-shot generations.

Deployment Guide

1. Clone the repo and create a Python virtual environment

bash

1git clone https://github.com/ResearAI/AutoFigure-Edit.git && cd AutoFigure-Edit && python -m venv .venv

2. Install dependencies and prepare environment variables (e.g., model provider API keys)

bash

1source .venv/bin/activate && pip install -r requirements.txt

3. Feed a method section as input and export an editable SVG figure

bash

1python main.py --input ./examples/method.txt --out ./outputs/figure.svg

Use Cases

Core Scene	Target Audience	Solution	Outcome
Method Figure Auto-Generation	Researchers and paper authors	Convert method sections into editable SVG diagrams with fast refinements	Compress figure work from hours to short, controlled iterations
Enterprise Whitepaper Illustrations	Solution Architects	Generate architecture diagrams from system descriptions with consistent styling	Higher doc consistency with less back-and-forth and rework
Teaching Slides as Visual Assets	Instructors and TAs	Turn lecture summaries into structured SVG figures for continuous reuse	Faster updates and clearer version control

AutoFigure-Edit

What is it?

Pain Points vs Innovation

Architecture Deep Dive

Deployment Guide

1. Clone the repo and create a Python virtual environment

2. Install dependencies and prepare environment variables (e.g., model provider API keys)

3. Feed a method section as input and export an editable SVG figure

Use Cases

Limitations & Gotchas

Frequently Asked Questions