Cross-AI
Cross-AI Team Program
Honest multi-runtime orchestration with sequential envelopes, conflict logs, and Human Yes hard-stops. Live packages: Research Desk and Operations Desk only.
Cross-AI vs Marketplace vs Library
Cross-AI
Orchestration of multiple honest runtimes (fetch, Ollama, file audit, coordinator, stop-gate). Sequential envelopes, conflict logs, Human Yes gates. Product + public proof — not a Library Verified badge.
Marketplace
Ready-to-use AI teams / paid desks buyers run or buy (Teams, AI team marketplace). Separate commercial surface. Cross-AI does not sell itself as a Marketplace listing.
Library (Build Library)
Free local recipes (Ollama Modelfiles, workflows). Untested until an operator marks tested. Cross-AI may use Library models as seats; it does not flip Verified.
Live now
LIVE Only Research Desk and Operations Desk are published.
Research Desk
Sequential research across public Library pages + local Proof Engine audit + executor synthesis. RESULT PASS · 2026-09-10 JST.
Operations Desk
Switchboard-style triage with stop-gate. External reply STOPPED at Human Yes and was NOT sent. RESULT PASS · SAMPLE ticket.
HOLD — not published
These packages are ranked HOLD and have no product pages live:
- Software Team — PARTIAL / Claude Code·CloudAgent UNAVAILABLE (not invented)
- Content Desk — design stub
- Trading Research Desk — design stub · ANALYSIS≠EXECUTION
Support-triage PE elevate remains HOLD. GrokHub ingest is NO-GO. No Verified flip from Cross-AI.
What this is not
- Not a Library Verified claim
- Not simultaneous multi-AI debate theater
- Not a send / mass-DM / X ops surface