That's a good question. We wanted to get as far away from just OpenAI wrapper as possible, that's why:

1. We use multiple different web-scraping scripts and APIs to get latest, most relevant data (especially for market and competitor analysis);

2. We use multiple different LLMs, including fine-tune of Mixtral, Perplexity and Gemini (and we chose LLMs based on their strengths for different tasks);

3. We added LLM techniques like RAG, chain-of-thought and others to get more accurate and less "hallucinated" results;

4. We give quantifiable feedback for analyses, instead of qualitative (that you would typically get from ChatGPT), so that it's easier to compare ideas and see up/downsides. We also tried to make these scores as consistent as possible (if you try to validate your idea with some LLM and ask it to give a score, it will just give you some random number every time you re-generate the message);

5. Finally, you can share your idea with the community in one click, or save them conveniently in your profile. You can find interesting ideas in community, or find similar-minded people and work together on your idea.

Apart from that, we have dozens of features planned for Yazero, which will make it even better platform for validating your ideas. Social media analysis, smoke tests, tiktok-like idea swiper to get quick validation, pitchdeck generation and so much more. If this app gets enough users, we will continue working on it and make it a lot better.