I was always stuck at the ideation phase when wanting to build or contribute to AI products.

I needed a clearer picture of the AI ecosystem — one that wasn't too low-level, but still had enough depth to actually be useful.                              

So I paid for multiple deep research agents to map the whole thing out, then cleaned up the results myself (with a little help from my agent bros Claude and Codex).               

The output: 5 layers of the AI stack — from consumer-facing products down to the APIs and tooling underneath — broken down into 200 domains and 2,500+ specific opportunities, all backed by hundreds of real sources.

You can drill down level by level until something clicks, or open the graph view to see the full picture at once. The goal is to help you find where in the stack you're best positioned — or just spark an idea you hadn't considered.             

This was the methodology behind it: Five of the most capable AI research agents — OpenAI Deep Research, Gemini Deep Research, Perplexity, Onyx, and CellCog — each independently analyzed the full AI landscape. No shared context, no cross-contamination between agents. Before any code was written, a structured patch was drafted per layer. The taxonomy reflects research consensus — not any single model's opinion.

This is why the focus areas feel less like idea slop. They come from real sources, validated across multiple independent runs.

Built it because I needed it. Sharing it because maybe you do too. It's Open Source and free, no signups, no newsletter. Use it and build great amazing AI products, let me know if you do!

If you want to contribute, fix a typo, or add opportunities I've missed, feel free to create an issue or PR: https://github.com/nc1z/ai-opportunities

Hope this helps! Pretty sure we're all scrolling for ideas anyway 😀