The Best of AI project has announced a transition to a fully open management model, open-sourcing its entire technology stack: from the website's source code to tool data. Instead of traditional databases, flat files are now used, and the content update process has been moved to GitHub.
What Happened
The Best of AI project has switched to using flat files in Markdown and YAML formats instead of classic databases (Postgres/CMS). The website is now statically generated using Hugo and deployed via GitHub Actions to GitHub Pages. Content and leaderboard management is handled through the GitHub Pull Request mechanism.
Context
Previously, managing data catalogs often required centralized systems and heavy databases. Transitioning to a Git-driven approach (Data-as-Code) allows Git to serve as the primary interface for data management, simplifying auditing and reducing infrastructure costs.
Why It Matters for the Industry
For the industry, this move demonstrates a trend toward lightweight and transparent data catalog management. Using a Git-centric architecture and an automated CI/CD process via GitHub Actions sets a precedent for creating decentralized and verifiable AI tool aggregators without the need to maintain complex server infrastructure.
Why It Matters for Users
Users gain the ability not only to consume information but also to actively participate in shaping rankings by suggesting new tools or correcting existing ones via Pull Requests. Additionally, the data is available in JSON format, making it easy to integrate into their own research pipelines or projects.
What Is Not Yet Known / Limitations
A shift in the risk profile: control focus moves from database administrators to repository moderators, which requires the development of new content validation mechanisms.
Sources
Author
Look at AI, Editorial Staff