Governed AI coding for regulated Swiss teams
Swiss-jurisdiction governed AI coding.
An integrated AI coding environment built for regulated software teams in banking, insurance, pharma, and healthcare. In active development, with first pilots targeted for H2 2026.
Available today
What works today
We are precise about what already runs. One capability is in daily use; everything else on this page is what we are building toward.
An integrated AI coding environment with model-agnostic routing
An IDE that routes across hosted and self-hosted open models, so a team is not tied to a single provider. The founders use it hands-on on real codebases.
Designed, in development
What we are building
Three layers are designed end-to-end, informed by 15 buyer interviews, and not yet implemented. They are intended to turn a capable AI IDE into something a regulated organisation can approve.
- Designed · in development
Governance & policy-enforcement
Designed to apply an organisation's policy to AI-assisted changes: what may run, where, and against which models. The aim is that AI coding fits existing approval processes.
- Designed · in development
Audit & traceability
Built for a defensible record of AI-assisted work: what was proposed, by which model, and what a person approved. Intended to support review and regulatory scrutiny.
- Designed · in development
Enterprise context
Our approach is to connect the environment to an organisation's own context: its codebases, standards, and systems. Designed to keep AI assistance relevant without sending that context outside an approved boundary.
Control boundary
Code and AI processing, designed to stay inside an approved boundary
Our approach is to keep source code and AI processing inside an approved, customer-controlled or Swiss-hosted environment, designed to stay within a boundary under the customer's control. It is the technical expression of governance and approvability, and today it is a design rather than a finished capability.
Request
Routed
Delivered in-boundary
Designed · in development
Approved, customer-controlled or Swiss-hosted environment
Why it matters
Built around approvability, not just where data sits
For teams under FINMA and the revised FADP, the hard part of adopting AI coding is not reaching capable models. It is being able to approve, govern, and account for how they are used. Our focus is that approval model for regulated Swiss software teams. Staying in-jurisdiction matters, but on its own it is not the story.
Roadmap
Where we are, honestly
An early-stage venture, building in the open about timing.
Targeted September 2026
Integrated end-to-end prototype
Targeted H2 2026
First pilots with design partners
The demonstration
The governance approach, shown once as a demonstration
A single scripted demonstration shows how the governance, audit, and context layers are intended to work together around an AI-assisted change. It was shown once, at an early pitch event. It illustrates the intent. It is not a running product, a pilot, or repeatable functionality.
Design partnership
Explore a design partnership
We are looking for a small number of regulated Swiss teams to shape the product with us ahead of first pilots in H2 2026. If that might be you, tell us a little about your context.
Prefer email? Write to hello@helveticode.dev.