About
About chataiopensource.com
Our Mission
chataiopensource.com is a self-hosted, open-source platform for comparing AI models through blind, anonymous battles. We believe that transparent, community-driven evaluation is essential for advancing AI technology responsibly.
How It Works
When you submit a prompt, two randomly selected models respond side-by-side without revealing their identities. You vote on which response is better (or if it's a tie), and these votes are used to calculate Elo ratings that power our leaderboard.
This blind testing methodology reduces bias and ensures that models are evaluated based on the quality of their responses rather than brand recognition or prior expectations.
Technology Stack
chataiopensource.com is built with simplicity and self-hosting in mind:
- Backend: PHP 8+ with PDO/SQLite database
- Frontend: Vanilla HTML, CSS, and JavaScript (no frameworks)
- Model APIs: OpenAI-compatible API calls to free-tier providers
- Rating System: Elo rating algorithm with K=32, starting at 1000
Privacy & Data
Your battles and votes are stored locally in your deployment's database. We do not track personal information or send your data to third-party analytics services. Prompts are sent only to the configured model providers for processing.
Open Source
This project is open source and available for self-hosting. You can customize it, add your own models, and run it on your own infrastructure. The code is transparent, and the evaluation methodology is open for scrutiny.
Support & Contributing
As a community-driven project, we welcome feedback, bug reports, and contributions. Whether you're a developer who wants to improve the code or a user who wants to help test models, your participation makes chataiopensource.com better for everyone.