Cruxi – AI copilot for FDA 510(k) submissions

Hey IH 👋

We’re building Cruxi, an AI copilot for FDA 510(k) submissions.

It’s a very niche product. If you’re not in medtech or regulatory, it might sound like complete alien language – and that’s totally fine. We’d love feedback mainly from people interested in the business / indie side rather than the medical details.

What problem we’re solving

Early-stage medtech teams hit a wall when they reach FDA 510(k):
– figuring out the right product code and regulation
– finding solid predicates
– understanding what testing/regulatory evidence is expected
– then formatting everything into eSTAR and submission-ready docs

Most teams either:
– burn months trying to figure it out themselves, or
– pay 5-figure consulting fees very early in their journey.

What Cruxi does

We built a chain of AI agents that:
– classify the device and map it to regulations/product codes
– surface relevant predicates, guidance, recalls, safety comms
– suggest likely testing and documentation expectations
– draft eSTAR-ready sections you can review and edit

It’s not meant to replace experienced regulatory consultants. It’s meant to reduce the chaos and give founders/RA teams a clear starting point and drafts to work from.

Why we’re posting here

We know this is super niche for Indie Hackers. So rather than “judge it as a consumer app,” we’d really appreciate feedback on:
– Does the positioning make sense as a bootstrapped/B2B SaaS?
– Any ideas on how you’d approach distribution for such a narrow niche?

If you are in medtech / RA / consulting and want to play with it, we’re happy to give you access and would love brutally honest (but kind 😅) feedback.

👉 Product: https://cruxi.ai
👉 Product Hunt launch (in 5 hours): https://www.producthunt.com/products/cruxi?comment=5030895

Happy to answer questions about the product, the tech stack (GCP + Vertex + MongoDB + a chain of agents), or the go-to-market side.