Proof
PilotOS builds itself
The best proof that PilotOS can build a business is that it is already building one — its own. Below are real pull requests (improvements to its own code) that PilotOS planned, wrote, and merged. This list is live and updates on its own.
- fix(api): enforce Atlas tenant scope on /api/atlas/decisions
- fix(web): enforce tenant scope on AdPilot/WebPilot wing mutations (PIL-172)
- fix(billing): preserve subscription id on invoice.payment_failed webhooks
- fix: Stripe billing webhooks blocked by operator middleware (401)
- feat(nexus): PIL-179 real Vercel Sandbox executor runner
- fix(web): clear CodePilot send idle timeout on unmount
- fix(ci): nightly self-build smoke — use github.token instead of under-scoped PAT
- Canonical home for openclaw hook scripts
- feat: Atlas governed memory contract (IEI-60..IEI-70)
- Preserve pilotos.dev marketing site (apps/site) under version control
- PIL-176: public marketing site (Phase 5.6a)
- PIL-177: live /api/proof endpoint (Phase 5.6b)
- feat: PIL-173 — Stripe billing integration
- feat: PIL-172 — multi-tenant isolation (RLS + per-tenant scoping)
- feat: PIL-171 — real cockpit auth + signup
- feat: PIL-168 — DocPilot + AppPilot first real capability
- feat: PIL-169 — real /console operator flows
- feat: PIL-167 — AdPilot + WebPilot operator-approved mutation paths
- feat: PIL-170 — Slack bridge handles channel @-mentions
- fix: exempt /api/slack/events from operator-auth middleware
Want it building for you next?
The same engine improving PilotOS can plan and build your business.