Company

About Rhone

Rhone builds infrastructure for AI-operated software.

Our work spans Vango, a server-driven full-stack Go framework; Rhone Code, a contract graph and proof workbench for shipping AI-assisted software safely; Rhone Cloud, the managed runtime for Vango and hosted development; and Rhone AI Gateway, the stateful execution layer for model calls, sessions, runs, routing, and observability.

Core thesis

AI will not only write software. It will operate software, maintain systems, and participate in durable production work.

Why Rhone exists

The next software stack needs stronger foundations.

Chat logs, screenshots, loose prompts, and unstructured code edits are not enough for systems that will be built, maintained, and operated by AI agents.

Rhone is building explicit contracts, server-owned state, generated rails, manual seams, provenance, reviewable proof, and hosted authority into one connected stack.

The goal is not to make agents look busy. The goal is to make them useful, bounded, inspectable, and capable of doing real work inside systems that can be understood, reviewed, and shipped.

What we believe

Software should be generated, authored, verified, and operated with the same source of truth.

Truth should be durable.

The important facts about a product should live in source, specs, contracts, artifacts, and proof, not in transient UI state or private model context.

Agents should be powerful, but bounded.

AI systems should be able to build, inspect, repair, and explain software. They should not fabricate authority, bypass review, or confuse local evidence with production readiness.

Generation belongs where structure is stable.

The right system is not raw codegen or a chatbot editor. It is a compiler-backed workflow where intent, implementation, verification, and deployment stay connected.

What we are building

A connected stack for AI-native software.

Each product has a clear boundary. Together they connect intent, implementation, runtime, and AI execution.

Direction

Rhone exists to make AI-native software development trustworthy.

Agents need real authority to do useful work. They also need boundaries, evidence, review, and systems that can say no.

That is the stack we are building.