Company
Canadian frontier intelligence.
DipoleML builds frontier AI in Canada and puts it within everyone's reach: the raven model family, DCode, and the infrastructure to run them.
Mission
Build frontier AI in Canada, and put it within everyone's reach.
The frontier should not be a place you visit through someone else's terms of service. We build it here, price it so it can be reached, and design it so it can be trusted.
Vision
A Canada that creates frontier intelligence, and shares the benefit widely.
A country with its own frontier capability sets its own terms: in its economy, its research, and its digital sovereignty.
Alignment
Built for AI for all.
Canada's federal AI agenda rests on access, sovereignty, and trust. Those are engineering requirements here, not slogans.
Access
A real free tier, pay-as-you-go credits with no subscription required, and pricing that starts at zero. A student, a startup, and a ministry should all reach the frontier.
Sovereignty
A Canadian company, operating its own gateway, billing, and account infrastructure. Your AI stack should not live entirely in someone else's jurisdiction.
Trust
We meter counts and tokens to run the service; we do not log or store prompt and completion content. PIPEDA and GDPR rights are written into our policies.
Our story
From a research habit to a launch date.
DipoleML started in 2022 as a research habit: reading the papers, running the experiments, and refusing to accept that frontier work only happens south of the border. It grew into a small, research-first company that builds what it learns.
On October 1, 2026, that work shipped: the raven model family went live across DCode, the API, and the raven web chat, engineered and served from Canada.
2022
The habit begins
Research-first, self-taught, Canadian to the core.
2026
The system comes together
Models, a gateway, an agent, and a plan: build the whole stack.
October 1, 2026
Launch
raven Flash, raven Max, and DCode 1.0 go live.