Using standard markdown files for the data model yields major benefits:
- Zero Database Latency: All pages are statically generated at build time. The site loads almost instantly because it's just static HTML and CSS served via CDN.
- Git-based Workflow: Updating the database is as simple as pushing a new markdown file or merging a PR.
- Robust Local Filtering: By compiling the frontmatter metadata into a single JSON index at build time, I can run ultra-fast client-side filtering and sorting without making any API requests.
Tackling the Filtering Engine
The biggest hurdle with static directories is handling complex multi-select filtering (e.g., finding a tool that has an API, offers a free tier, and supports image-to-video generation).
Instead of relying on heavy client-side state libraries, I built a lightweight search index utility that parses the local JSON list on the client side using search parameters. The state is serialized directly into the URL, making every search filter deep-linkable.
What's Next?
The goal is to keep this repository of AI video production software as clean, fast, and structured as possible. I want to build a public GitHub repository where developers can submit pull requests to add new tools or update pricing directly, avoiding the need for a complex admin portal.
Check out the live version at findaivideo.com and let me know your thoughts on this static markdown approach! How do you handle directory structures in your own side projects?
Top comments (0)