AutoGEO

[ICLR'26] AutoGEO: a Generative Engine Optimization framework to automatically learn generative engine preferences, and rewrite web contents for more traction.

What is AutoGEO?

[ICLR'26] AutoGEO: a Generative Engine Optimization framework to automatically learn generative engine preferences, and rewrite web contents for more traction.

Overview of AutoGEO

What it is

AutoGEO is a framework for Automatic Generative Engine Optimization (GEO) that helps web content gain higher visibility in LLM-generated answers. Its paper has been accepted by ICLR 2026. The framework automatically extracts content preference rules from generative engines and rewrites documents to maximize visibility while preserving accuracy. It includes three core components: Rule Extraction, AutoGEOAPI (a prompt-based GEO model using extracted rules), and AutoGEOMini (a cost-effective GEO model trained with reinforcement learning). Evaluation metrics are GEO score (visibility) and GEU score (utility). Rule extraction is tailored to specific generative engines and datasets/domains; switching to a different engine or dataset/domain requires rerunning rule extraction for AutoGEOAPI and retraining AutoGEOMini accordingly.

Key features

  • Automatically mines content preferences from generative engines via Rule Extraction.
  • AutoGEOAPI: prompt-based rewriting using extracted rules.
  • AutoGEOMini: cost-effective model trained with reinforcement learning (SFT + GRPO).
  • Supports multiple datasets: Researchy-GEO, E-commerce, GEO-Bench.
  • Supports multiple generative engines: Gemini, GPT, Claude.
  • Unified evaluation framework for vanilla, autogeo_api, and autogeo_mini models.

Use cases

  • Optimizing web content to appear more frequently and prominently in generative search engine responses.
  • Improving visibility of e-commerce product pages, academic articles, or benchmark entries in LLM-generated answers.
  • Automating the rewriting of documents for better traction without human manual tuning.

Getting started

bash # Clone the repository git clone --recursive https://github.com/cxcscmu/AutoGEO cd AutoGEO # Run installation script bash install.sh # Activate environment conda activate autogeo # Configure API keys (required) nano keys.env # Add your API keys For training AutoGEOMini, additionally run: bash conda activate autogeo bash install_mini.sh See the repository README for detailed usage examples.

License and pricing

The project is licensed under MIT. It is open source with no stated pricing.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/cxcscmuautogeo.svg)](https://yunodb.com/tools/cxcscmuautogeo?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/cxcscmuautogeo?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/cxcscmuautogeo.svg" alt="Featured on YunoDB" height="20"></a>