Astartis x Codex - evidence-led developer security control plane.

I'm inspired by enterprise-grade software. Working part-time in IT internships while at university, I've seen how efficient and fault-tolerant enterprise networking and cybersecurity systems can be. But these systems are only affordable for large enterprises and government ministries in Botswana — small companies and solo developers can't afford things like Essl access-level door lock systems or Dell's write-once-read-many (WORM) server architecture, which gives absolute immutability. So I built a hyper-scalable system and used Codex/GPT-5.6 to debug its C++-heavy errors. It can run on a server or your own laptop by changing configs via my GitHub repo — making it cheap and reliable for regions like Botswana, where cloud hosting is expensive and infrastructure can be unreliable.

Congrats on shipping this. The local-first, disconnected-environment angle is the right call, most enterprise security tooling assumes constant connectivity and a budget most small teams don't have. Genuine question: how does evidence stay trustworthy once a device has been offline for a while and comes back online, do you reconcile state automatically or flag it for a human to review first?

Astartis x Codex

Enterprise developer security control plane around evidence

Enterprise developer security control plane around evidence