Builder
at heart.
Ratify Protocol: verifiable authority for AI agents.
Signed delegation proofs that let receiving systems verify what an agent is authorized to do before the action runs.
ratifyprotocol.com →And I work with select teams as a fractional CTO when available.
What I'm Building_
Ratify Protocol is my primary startup. Alongside it, I keep shipping tools and products that are fun, rewarding, and useful — from agent infrastructure to creative experiments.
Ratify Protocol
AI agents are getting access to real systems, but the systems receiving their requests can't verify what an agent is actually allowed to do.
Signed delegation proofs that let receiving systems verify an agent's authority before executing the action. Scoped, expiring, and checkable across company and runtime boundaries.
Deskhands
Running several coding agents at once means juggling terminal tabs, losing track of what each one is doing, and hoping none of them pushes something it shouldn't.
A Mac app where Claude Code, Codex, Gemini CLI, Antigravity and Grok work side by side in one small office. See every agent's live status, open its real terminal, and approve risky commands like git push before they run. Everything stays on your machine.
Fruit Fly Airlines
Connectome research is some of the most remarkable science of our time, and almost nobody gets to experience it.
Fly a real fly brain around the world. Pilot a fruit fly driven by its actual 165,122-neuron connectome through 166 cities with live weather.
Everything Built
Valuable connections fail to materialize because life gets busy — someone meant to make an intro, but the moment passed.
Aidan is your AI chief of staff for relationships. Tell it your goals and who you want to meet — Aidan reaches out, gets consent, and schedules the connection. No cold messages, no follow-up friction.
Crayfish
AI assistants are locked behind subscriptions, harvest your data, and can't actually do anything useful.
A personal AI that runs on your own hardware — manages email, remembers what matters, and gets smarter the more you use it. Free, private, open source.
Syncline
Scheduling meetings programmatically is a nightmare of OAuth, timezones, and edge cases.
One API call. Pure magic. AI-first scheduling that handles the complexity for you.
Synod
Single-model AI coding is limited by blind spots and biases.
Intelligent multi-model coding agent with adversarial debate. Where AI models debate, so you get better code.
VibePlay
Making music requires years of training and expensive instruments.
Turn gestures into music. The world's first AI Orchestra — anyone can be a conductor.
Still
Digital distraction is destroying our ability to read — 47% of articles are abandoned mid-scroll.
RSVP speed reading that lets you read 2-3x faster. Still your eyes, let the words come to you.
Recent Writing
Most of my writing is about the missing layer in agentic AI: how systems verify what an agent is actually authorized to do before it acts.
When Your Assistant Asks Mine
When personal assistants coordinate, each side must know what its own agent may ask or disclose.
What Trust Should Feel Like
As AI enters everyday institutions, people should be able to understand what they allowed and receiving systems should be able to check it.
When Machines Share Our World
As autonomous machines begin working alongside people, a receiving machine must know who authorized a request and whether it can safely act in its present surroundings.
Jev Can Choose the Tool. Who Authorizes the Action?
Typed model decisions can reshape agent loops, but a decision is still not authority. The receiver needs evidence it can verify before the side effect occurs.
Working Together
When I have capacity, I work with a small number of teams on a fractional basis — bringing the same hands-on building I do on my own products.
FRACTIONAL_CTO
Technical Leadership
Hands-on fractional CTO support for product, engineering, architecture, vendor selection, and build-vs-buy decisions.
AGENTIC_SYSTEMS
Strategy & Architecture
Designing agentic systems that hold up in production — what agents should do, what they should never do, and how to verify the difference.
PRODUCT_BUILD
From Idea to Shipped
AI-first product development with rapid prototyping and iterative shipping, built alongside your team.
About Me
I'm a builder at heart. My primary focus is Ratify Protocol, the startup I founded to give AI agents verifiable, scoped authority, so the systems they touch can check what an agent is allowed to do before it acts.
Kekwanu Labs is where everything else lives: the tools and products I keep shipping because they're fun, rewarding, and useful, like Deskhands, Synod, and Crayfish. When I have capacity, I also work with a small number of teams as a fractional CTO.
"Ship fast, learn faster. The best strategy emerges from building, not planning. AI changes everything — adapt or get left behind."
Products, prototypes, and production systems. From idea to shipped.
Fractional CTO support for select teams: architecture, agentic systems, and build decisions.
Essays on agent authorization and what it takes to trust AI that acts.
Open source tools and experiments pushing the frontier of AI.
Start a Conversation
Want to talk about Ratify, something I've built, or a fractional engagement? I'd like to hear what you're working on.
Direct line for inquiries
[email protected]SCHEDULE
Book a discovery call
calendly.com/kekwanulabsConnect and message
/in/chuksonwunemeHave a Team or Project in Mind?
When I have capacity, I work with a small number of founders and product teams as a fractional CTO. No generic sales pitch, just a working conversation.
SEND_INQUIRY