App's mission is to create source of truth for COMPLETE and LIVE set requirements of the product, which are tracable to the raw communication history, adjacent requirements, tasks, tests and the code.
Today's gap in autonomous coders, that agents do not share any of your assumtions and the way how you map your codebase with words, concepts and priorities. And in the same time building full requirements with existing tools is not realistic effort anywhere other than in safety critical verticals.
First user flow available is to interview you as main stakeholder, digest specifications, diagrams and autocode clickable web app preview. Then you can redline and editing specs, by selecting doc / or atomic requirement. Then you can export markdown, PDF, or DOCX. (no full throttle vibe coding of the web preview yet! Also diagrams editing is not available for now)
Initial interview digesting uses Gemini Pro 2.5 and might take 2-3 minutes for completion. So give it its time!
I used to build a drones for a living as a founder, guidance-navigation-control and system engineer. Nautex AI is my effort for productizing full stack system engineering activities without its pre-LLM era costs.
Next week milestone is to bring MCP server, so the Cursor/Windsurf can use it.
Please share any of your feedback!
sneha_tamal•9mo ago
how well does it handle changes in product direction mid-build? Like if priorities shift, can it re-align requirements easily?
Homo__Ludens•9mo ago
It is designed to be fully LLM editable with full control over changes. So I target the use case when any change requested calculate whole chain of the impact for the review.
Right now it can rewrite the document in a specific way on follow-up promts. This rewriting uses gpt4o tool calls (so kind of basic reasoning), over the weeked will put Gemini Pro 2.5 to that editing service (then going to try grok 3). My expectation is that its going to feel like your doc got edited by chief architect (gemini pro) vs junior pm (gpt4o)
Homo__Ludens•9mo ago