May 7 2026
We indexed every Act, regulation and judgment in Mauritius
1
WHAT WE BELIEVE
We started by building the hardest thing first.
The obvious first product for an AI company in Mauritius would have been something simple. We built an AI trained on the entire published corpus of Mauritian law instead, because that was the problem worth solving, and because doing it properly would tell us whether we could do anything else.
That decision set how we work. We build products rather than assemble integrations, and we stay in one jurisdiction deeply enough to be genuinely useful in it, Mauritian law, Mauritian regulators, and the confidentiality duties Mauritian professionals already owe their clients. The same team now applies that to finance, automation and security. The practice areas are separate on paper, but they share one discipline: know where client data travels, prove where an answer came from, and leave the firm able to run the system without us. That is what we mean by building AI for firms that cannot afford to get it wrong.
2
Our team
You work with the people who built Themis. There is no account management layer.
Builders, not resellers
We build products, not just integrations. Themis required assembling the complete published corpus of Mauritian law: 4,100+ Acts, 57,000+ judgments and 17,800+ regulations. Most of what we do for clients is far simpler than that, which is exactly the point.
Based in Saint-Pierre
Saint-Pierre, Mauritius. We work with clients across the island, and we come to your office, the Map stage requires it.
Traceable by design
Every output that matters should be traceable to a source, and security is a design input rather than a later phase. The measure of a project is whether the firm still uses it in a year.
3
Our driving forces
Knowledge as a Driver
We built an AI trained on the entire published corpus of Mauritian law before we built anything simpler. Depth in one domain beats breadth across many, and we invest in understanding how our clients actually work rather than generalising about it.
Risk Management
We run an offensive security practice, so we know what an attacker does with a badly configured AI integration. Every system is threat-modelled, access-scoped and logged before it touches client data.
Operational Excellence
We work in two-week sprints and finish by training the people who will use the system. The measure of a project is whether the firm still relies on it a year later.