Platforms
Juego builds and deploys software platforms that serve as the chassis for product engineering. JaiCode runs the gated agentic SDLC. Boutique and Box are how institutions get products on that chassis. FDEs put it in the field.
Stack
Operating system first. Product offerings on top. Forward Deployment Engineers put them in the field.
01 · Operating system
Local-first IDE · CLI · companion for Map → Operate.
Modes constrain tools; plans unlock side effects; Agent Review scores the diff. BYOK for Claude, OpenAI, Gemini, and private endpoints. Context packs and path grants keep GenAI inside safe bounds—especially in regulated domains.
02 · Product offering
Bespoke GenAI products, crafted end to end.
Greenfield and high-craft builds where FDEs and your owners share one JaiCode workspace—from brief through MLP with Plan unlock and Agent Review on every generative cut.
03 · Product offering
Packaged agentic delivery your team can run.
A packaged path to stand up gated GenAI delivery—modes, context packs, and review—so your engineers keep the cadence after FDEs step back.
JaiCode factory
Interactive capability line Forward Deployment Engineers run on the platform.
Delivery factory
JaiCode is the operating system. Generative AI is the engine; context engineering and gates are the line. Intent enters on the left; FDEs advance work through Map → Operate with Claude, OpenAI, Gemini, and other BYOK models. Click any station for detail.
Overhead · Developer experience
In
Out
Intent moving ↓
Shipped software
In
Intent · legacy
Out
Shipped software
Now on the line → Map: Understand what exists
Factory floor · Foundations
In the room
Screenshots from the JaiCode workspace—IDE, Agents Window, Agent Review, and source control.
Desktop IDE
Gates
The platform does not let agents free-fire into main. Modes constrain tools; plans unlock side effects; Agent Review scores the diff.
Ask
Read-safe search, LSP, research, and exploration—no project writes. Where mapping and brownfield discovery start.
Plan
Research the repo, clarify scope, lock a structured plan. Humans approve before Build.
Debug
Evidence-first investigation, then targeted fixes when the diagnosis is solid.
Build
Edits, documents, shell, browser, MCP, and checkpoints when the plan is unlocked.
Agent Review runs against the diff—findings and a verdict before merge. FDEs and client leads disposition pass, fix, or escalate. No silent generative merges.
Trust
Agents run on the machine against your checkout. Generative work stays inspectable—not trapped in a vendor sandbox you cannot audit.
Bring Anthropic Claude, OpenAI, Google Gemini, or other approved keys under your procurement. Usage stays attributable in the companion.
Path grants and run policy scope what agents may read and execute—especially critical in regulated domains.
Invite-only accounts, device approval, and admin. Your org owns seats; FDEs join as collaborators.