HR Workmates
Onboarding plans, policy answers, and people-ops drafts — ready for a manager’s review.
AI Studio · Building software for the age of AI agents
An AI studio for what works alongside AI.
Most of the value in AI lives in the layer above the model — the teammates, the rails, and the surfaces. We incubate the software that lives there. Quietly useful tools that make the rest of your AI stack better.
What we build
Three pillars. One thesis.
01 · Teammates
AI Workmates
Software utilities that act as extended teammates alongside your AI agents. Memory, coordination, role-aware behavior — the bits that turn a single LLM into a team.
Explore Workmates02 · Rails
AI Infrastructure
The plumbing for serious AI work. MCP servers, command-line tools, adapters — the gear that makes agentic systems run.
Explore Infrastructure03 · Surfaces
AI Experience
The human side of AI. Small, well-designed tools that smooth the rough edges of how people actually use AI every day.
Explore ExperienceHow the studio works
Foundation models are remarkable. They are also incomplete on their own.
They need teammates to extend their reach, rails to keep them honest, and a designed surface for the humans on the other side of the conversation. That space is where the next decade of useful software gets built — and where the studio lives.
We are a small team with a long thesis. We notice gaps, build sharp tools to fill them, ship to real users early, and keep refining what works.
01
We notice a gap in the stack.
02
We build a sharp tool to fill it.
03
We ship and watch it work.
04
We improve it continuously.
Selected projects
A glimpse of what is in the works.
IT Workmates
Triages and resolves the long tail of help-desk work; escalates only what truly needs a human.
MCP Gateway
A secure and scalable MCP gateway with built-in observability and governance.
Cost Control
Contextual cost auditing and budgeting that keeps AI expenses in check while letting your agents run at full tilt.
Continuous Context
Extend context beyond context windows and optimize the context to specific outcomes and goals.
Prompt Palette
Prompt generators based on specific job roles and personas that evolve with context.
From the studio
Notes on building alongside AI.
Workmates
The workmate is not the agent.
A short framing essay on why the persistent, coordination-aware layer above an agent deserves its own name — and its own software.
InfrastructureMCP in plain language.
What the Model Context Protocol actually is, why it matters for builders, and the things it quietly fixes.
Quietly hiring builders.
We are growing a small team of people who care about the unglamorous parts of AI. If that is you, we want to hear from you.