Local-first AI infrastructure that checks itself as it ages.
The studio behind Sextant, a memory and orchestration platform that re-checks its own stored facts as they age. Vendor-neutral, reproducible, local by default. No cloud dependency required, no per-token cost to run.
A few of the things, with more on the way.
Sextant, public on GitHub, and the research library behind its guides. The terminal suite and the extracted libraries follow as each is ready; the Work page has the shelf.
Sextant
Keeps an LLM's long-term memory correct through epistemic decay and scheduled re-verification, and orchestrates autonomous work through a task queue, cron scheduler, quality gates, and model-escalation routing.
Research library
4,000+ reports, nearly all carrying a current verification review, re-checked on a staleness schedule, with a citation guard that rejects fabricated sources.
The same studio runs a small, free publication for everyday users: plain-English writing about the technology that actually affects your privacy and data, printable scam guides. The first guides are in preparation; those pages are layout previews until then.
Direct is fine. Skip the ceremony.
Collaboration, infrastructure work, or questions about Sextant.