The rule on this page: If it's not done or actively being built, it says "Planned" — not "Coming Soon." We don't ship promises.
The foundation is running. Local agents, local inference, zero cloud dependency for core operations.
Multi-agent system (Pixie, Shield, Matrix, Thinker, Tonka, Journey, Ambassador, Justice)
Local LLM inference via Ollama — $0 per query
Autonomous task board — agents pick up and execute tasks without manual prompting
SmartQuoteAI — job quoting and estimating for trades and contractors
Budget, Calendar, MindVault, LionGate Travels, Rows
Telegram integration — agents report to you daily
liongateos.com — 100/100 Lighthouse performance score
Making the system ready for more than one user. Right now everything runs on one machine for one owner. This phase adds the separation layer.
User accounts with isolated task boards — your agents stay yours
Row-level security — no user can see another user's data
Auth-gated Mission Control — login shows only your agents and tasks
Right now Mission Control points at one machine. This phase lets you point it at any node — including your own hardware.
User-configurable gateway URL — connect Mission Control to your own server
"Connect Your Node" setup screen with live connectivity test
Self-hosting documentation for technical users
A one-command installer. You run it, you get a personal agent node on your own machine. Your data never leaves your hardware.
Single install script — works on Ubuntu/Pop!_OS/Debian
Choose your first agent: Quoter, Scout, or Assistant
No GPU required for basic tier — runs on CPU with a lightweight model
Auto-connects to Mission Control on first run
The full vision. Users don't use our agents — they run their own. A private Quoter for a contractor in Calgary. A Travel Scout for a solo traveler in Lisbon. All running locally, all coordinated through Mission Control.
Remote task handoff — dispatch tasks to agents on any machine
Encrypted task payloads in transit
Per-user agent roster and usage dashboard
Free tier: 1 agent, CPU-only. Paid tier: full roster, GPU inference
What We're Not Building
Not a cloud platform. We're not pivoting to SaaS. The local-first architecture is the product, not a constraint.
Not a ChatGPT wrapper. Every agent runs on open-weight models you control. No API key required to use core features.
Not enterprise software. Built for one person running a trade business or small operation — not IT departments.
Request a Feature
If you're a contractor, builder, or small business owner and something is missing — tell us. We read every submission. No spam, no newsletter, just the request.
Got it — thank you. Antonio reviews these personally.