Canopy Ultimate — Three AI Engines, 40+ Agents, Full Company Stack

# Canopy Ultimate The premium Canopy deployment. Everything in Command Center, plus OVERSEER routing, GUARDIAN pre-execution budget gate, three independent AI adapters with their own token budgets, and SENTINEL enforcer with per-adapter Telegram alerts. ## What You Get - **OVERSEER** — routes every task to the right engine and team - **GUARDIAN proxy** — pre-execution budget check on every tool call - **FORGE** (claude-sonnet-4-6) — reasoning, architecture, multi-file tasks - **CODEX** (gpt-4o) — generation, content, high-volume work - **OpenRouter fallback** — claude-3-haiku for overflow and burst - **Per-adapter token budgets** — 3M/day FORGE, 3M/day CODEX, 1M/day fallback - **SENTINEL** — Telegram alerts at 50%, 80%, 95% per adapter; hard stop at 100% - **CRUCIBLE** — security auditor embedded in the agent roster - **40+ named agents** — 7 departments: Engineering, Product, Sales, Marketing, Ops, CS, Executive - **cognitive-os workspace** — full company structure seeded on first login - **Full server security** — nginx TLS 1.3, UFW, fail2ban, SSH hardening - **Daily backups** — PostgreSQL + Redis + config ## Stack Elixir/Phoenix backend · PostgreSQL 16 · Redis 7 · SvelteKit 3D frontend · nginx · Docker Compose ## Requirements - Ubuntu 22.04 VPS — 2 vCPU, 4 GB RAM minimum ($24–48/mo) - Domain name with DNS control - Anthropic API key + OpenAI API key (both required for dual-engine) ## Time to deploy: 20 minutes