VIBIN AI LABS

An evolving frontier

We are on a journey to unbounded recursive superintelligence

Read the research agenda See the products

Products are where our research meets the world. Research is where our products come from.

We think the loop between the two is what shapes self-improvement research

01 · Applied AI Products

Agents doing real work, for people who are not AI experts.

Two products, two domains. Each one is a long-running research deployment we learn from.

Cadex · Web platform

One intelligent workspace for everything happening on the ground.

Built for land, infrastructure and environmental projects. Everything the project knows, in one place, with an assistant that understands all of it.

Visit Cadex

Vibin · iOS app

An agentic investing app that teaches you to invest.

Beginners learn with an agent that remembers their plan, then graduate to paper and broker-connected trading. 

02 · Superintelligence Research

We study recursive self-improvement both with and without a boundary.

Direction A

Bounded Meta-RSI Systems

Self-improving systems with an explicit envelope. A meta-level improver edits a task agent's capabilities and its own improvement procedure, but only inside a fixed edit surface, checked by an independent verifier, and measured on benchmarks across domains.

Direction B

Open-ended RSI Systems

Open-ended search over agents, capabilities and improvement procedures: systems that keep generating novel problems and solutions without a fixed objective.

03 · ABOUT

Things about us

VIBIN AI LABS is incorporated in British Columbia, Canada. We are a small team of researchers and engineers who believe the next generation of labs will ship products and publish research from the same loop.

We work in the open where we can: our products are used by real people, our benchmarks and design documents are written to be read, and our research directions are stated plainly so they can be challenged.

Contact

info@vibin.ca