There’s a pattern we call cognitive drift: AI agents lose context between sessions and re-discover what you already decided. You brief Monday’s agent, then brief Wednesday’s agent the same way. Decisions evaporate. Quality varies. The human becomes the bottleneck — not because the tools are slow, but because they forget.
Enginehaus started as a question about whether you could fix this structurally rather than through better prompting. The answer, it turns out, is yes — but the interesting part is what happens next.
What we’re learning
When decisions persist and context compounds across sessions, the nature of human-AI collaboration changes. You stop managing agents and start coordinating with them. The work shifts from instruction to intent.
The key insight has been that structure beats instruction. Agents forget what you tell them; they can’t skip code. If something matters, it needs to be a gate — not a suggestion in a prompt. This is the principle Enginehaus is built on, and it’s held up across every project we’ve tested it on.
Decision archaeology emerged from this: every architectural choice, tradeoff, and “why not” gets logged with reasoning and becomes retrievable by future agents. We didn’t plan this as a feature — it fell out of the observation that the decisions you made last month should inform the work that happens tomorrow.
Beyond software
The original experiment was for software engineering. But the coordination problem — context that evaporates between sessions — turns out to exist everywhere people work with AI. Writing, research, strategic planning. Enginehaus now supports domain profiles: configurable phases, decision categories, and quality gates for different kinds of work. These are seeds, not finished products. Profiles are JSON files with no code required — designed so anyone can create one for their domain.
Where it is now
Built on the Model Context Protocol. Works with Claude Desktop, Claude Code, Cursor, Windsurf, Gemini CLI, Kiro, and any MCP client. Runs locally — your data stays on your machine. A learning engine aggregates patterns across projects.
Enginehaus is the technical foundation for the other Applied Humanism experiments — the coordination layer they build on.
npx enginehaus init