Why Canada needs its own frontier AI lab

Frontier AI means models at the leading edge of what the field can do: models that write and fix real code, read a long contract in one pass, and plan work over many steps. Most of that work happens inside a few very large companies, nearly all of them based outside Canada. Dipole ML Inc. is a Toronto-based Canadian frontier AI lab that builds the Raven model family, DCode, and the DipoleML API. Here is why we think Canada should have labs of its own, and what being a Canadian AI company means for you.

What is a frontier AI lab?

A frontier AI lab builds models at the leading edge of the field, and the products people use them through. That takes more than a model. It takes the research behind it, the engineering that makes it dependable, and the judgment that comes from watching how people really use it. Those skills build on each other. The places that do this work keep getting better at it, and the people who learn it tend to stay where the work is.

Why should Canada build its own?

Canada has been central to modern AI. Much of the early research on deep learning was done here, and Canadian universities still train some of the best people in the field. Too often, though, the companies built on that work, and the jobs, revenue and experience that come with them, have grown up somewhere else.

A Canadian frontier lab keeps more of that at home: the skills, the decisions about how frontier models are built, and the companies that grow around them. It gives people and organizations here a capable choice from a company that answers to Canadian law. And it shows that a Canadian company can compete with the best in the world on the quality of its work.

What does a Canadian AI company mean for you?

We are a federally incorporated Canadian company based in Ontario, so your personal information is handled under Canadian privacy law. Under PIPEDA you can ask to see what we hold about you, ask us to correct it, and, if we do not resolve a concern, complain to the Office of the Privacy Commissioner of Canada.

Our Privacy Policy spells out the rest in plain language. When you are signed in, your chats are saved so they are there on every device. Your content is encrypted at rest and used to develop, improve and train our own models. It is never sold, never used for advertising, and never used to train models for anyone else. Access to stored content is limited to the team members who need it to operate and improve the service. You can delete any chat at any time, and when you delete your account, your chats are removed within 30 days. How we handle your data, step by step, is in How we protect your data.

What else do we stand for?

Two more principles shape how we build. The first is access. Raven has a real free tier whose limits refill on their own, paid plans start at $20 a month, and the API is pay as you go, with every price published on our pricing page. The second is autonomy. We are independent and bootstrapped, with no outside venture capital, and the DipoleML API is a drop-in, so moving to it, or away from it, does not mean rebuilding your product.

Does this mean you should only use Canadian AI?

No. Use whatever serves you best. Our aim is that Canadians, and everyone else, have a strong choice built by a Canadian company: priced so it can be reached, and backed by a privacy policy you can read and hold us to. That is what we mean by Canadian frontier AI, available globally. The company story is on the About page.