Building the future of
artificial agency
A research lab and infrastructure company building systems where intelligence grows with people through memory, co‑agents, and reinforcement.
Manifesto
Intelligence Won’t Arrive as a Finished Product; it Will Emerge
Agency creates value. Left un-checked, that value concentrates behind APIs and monopolies. We bend the curve toward participation so everyone can build
it—and build on it. Julep is our first step: primitives for memory, co-agency, and reinforcement, with study mode for research and production mode for deployment. Next up: a Memory MCP for editing agent memory, an Agent Store for forking behaviors, and deeper reinforcement hooks.
Julep is our first step: primitives for memory, co-agency, and reinforcement, with study mode for research and production mode for deployment. Next up: a Memory MCP for editing agent memory, an Agent Store for forking behaviors, and deeper reinforcement hooks.
We refuse enclosure posing as safety. We choose to publish, codify, and invite critique. If you’re building the same future, come help us test, argue, improve, and deploy. The future of intelligence will not be delivered by one model; it will be made—together.