Built from real problems.
Not from a whiteboard.

Traverse is personal research turned into a runtime. It exists because the same engineering wall kept appearing in different companies, different industries, different codebases. This is the answer to that wall.

Years building products that kept hitting the same wall.

Enrico Piovesan has spent years shipping products across travel platforms, payments infrastructure, automotive software, and design tools at global scale. The companies were different. The tech stacks were different. The wall was always the same.

Business logic gets written once for the environment it needs to work in right now. Then the next environment comes. A mobile app. An edge deployment. An AI pipeline. The same logic gets rewritten. Sometimes by the same team. Often not. The second version drifts from the first. The third is unrecognizable.

Teams that were there at the start could navigate it. Teams that arrived later could not. People left. The codebase became archaeology.

Traverse is what happens when you get tired of watching that pattern repeat.

Software should work for users, not for the runtime it happens to run on.

Enrico Piovesan

The questions came first. The runtime came last.

Traverse did not start as a runtime idea. It started as a series of questions that each had to be answered before the next one could be asked. Each question became a published piece of work. The runtime is where those answers become executable.

2023
Client-Side Microservices Architecture

Asked why the browser was treated as a thin presentation layer when it was already a capable execution environment. Showed how microservice patterns could work entirely client-side, without a server in the loop for business logic.

2024
Universal Microservices Architecture (UMA)

Asked why business logic had to be rewritten for every runtime. Defined the architectural pattern for capabilities that run anywhere. Published as a book on Amazon: 13 chapters, with runnable Rust and WASM examples throughout.

Read on Amazon → · universalmicroservices.com →
2025
Contract-Driven AI Development (C-DAD)

Asked why AI agents struggled to navigate codebases that had no contracts. Defined a methodology where every capability exposes a machine-readable contract that agents can read and reason about. Published as a whitepaper.

Read the whitepaper →
2025
Traverse

The runtime where the answers to those questions become executable. WASM-based, contract-governed, trace-producing. Runs business logic in browser, edge, cloud, and AI pipelines without rewriting it. Open source, Apache 2.0.

View on GitHub →

Shipping. Honest about what is not done yet.

v0.7.0 is out. The core runtime, CLI, browser adapter, MCP server, and React demo all work. 9 governing specs. 100% test coverage on what ships.

Some executor targets are planned but not yet shipped. Edge and cloud adapters are next. The Python SDK is experimental and on the roadmap. The public roadmap is honest about what is shipping now and what comes later.

This is one person building something real. Not a startup with a deck. Not a proof of concept that will never ship. The code is there. The specs are there. You can read both.