Roadmap — built in the open.
What we are doing, in order, and what “done” actually means. Dates are targets, not promises; a phase only moves when its work is tested.
| Phase | Scope | Status |
|---|---|---|
| Phase 1 — Foundation | Competitor research, name and brand, repositories, landing page, registration forms, free infrastructure, documentation. | In progress — Nov 2026 |
| Phase 2 — First AI model | Free GPU environment, run a suitable open-weight model, benchmark inference, basic chat interface, API prototype. | Planned |
| Phase 3 — GPU agent | Hardware detection, GPU benchmark, contributor dashboard, secure registration, test machines connected to the coordinator. | Planned |
| Phase 4 — Marketplace | Scheduler, approved workers, routed test inference requests, reliability measurement, actual operating costs. | Planned |
| Phase 5 — Community | Discord launch, development updates, recruit initial GPU owners and developers, gather feedback. | Planned |
| Phase 6 — Private beta | Targets (goals, not achievements): 20 functioning GPU contributors, 10 developer testers, 3 supported models, working API, public benchmark results. | Planned |
| Phase 7 — Public launch | Only when security testing is complete, the API works reliably, provider consent is implemented, pricing is economically viable, and legal/payment requirements are addressed. | Planned |
Private-beta numbers are targets we aim for, not results. Nothing here is presented as already achieved.