Canadian Sovereign Infrastructure

AI inference
that stays
in Canada.

One OpenAI-compatible API. 100+ models. Canadian data residency, PIPEDA compliance, and intelligent routing — built for the government and enterprise buyers who can't afford to route their prompts through Seattle.

Data residency 100% Canada
Models 100+
Compliance PIPEDA, SOC2
OPENAI ANTHROPIC MISTRAL DEEPSEEK COHERE GEMINI CANADA
Get started in 60 seconds

Drop it into your existing code.

One API key. One endpoint. 100+ models. Change the base URL and go.

1 Get API Key
2 Make a request
3 Scale
Set your API key
export NORTHGATE_API_KEY="ngk_your_key_here"
Base URL: https://northgate-ai.polsia.app/v1
# pip install openai
from openai import OpenAI

client = OpenAI(
    api_key="ngk_your_key_here",
    base_url="https://northgate-ai.polsia.app/v1",
)

response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Hello from Canada"}],
    max_tokens=100,
)
print(response.choices[0].message.content)
// npm install openai
import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'ngk_your_key_here',
  baseURL: 'https://northgate-ai.polsia.app/v1',
});

const res = await client.chat.completions.create({
  model: 'gpt-4o-mini',
  messages: [{ role: 'user', content: 'Hello from Canada' }],
  max_tokens: 100,
});
console.log(res.choices[0].message.content);
# Routes automatically to cheapest available provider
curl https://northgate-ai.polsia.app/v1/chat/completions \
  -H "Authorization: Bearer $NORTHGATE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [{"role": "user", "content": "Hello from Canada"}],
    "max_tokens": 100
  }'
# pip install northgate-sdk
from northgate import NorthgateClient

client = NorthgateClient(api_key="ngk_your_key_here")

response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Hello from Canada"}],
)
print(response["choices"][0]["message"]["content"])
// npm install northgate-sdk const { NorthgateClient } = require('northgate-sdk'); const client = new NorthgateClient({ apiKey: 'ngk_your_key_here' }); const res = await client.chat.completions.create({ model: 'gpt-4o-mini', messages: [{ role: 'user', content: 'Hello from Canada' }], }); console.log(res.choices[0].message.content);
Request
{ "model": "gpt-4o-mini", "messages": [ { "role": "user", "content": "Hello from Canada" } ], "route": "cheapest" }
route=cheapest uses Groq (free tier) by default when available
Response
{ "id": "chatcmpl-...", "model": "llama-3.3-70b-versatile", "choices": [{ "message": { "role": "assistant", "content": "Hello! ..." } }], "usage": { "prompt_tokens": 20, "completion_tokens": 18 } }
Model swapped automatically — your code stays the same
100+ models via single endpoint
Automatic fallback — never fails silently
PIPEDA-aligned audit logs on every request
route=cheapest | route=fastest | route=best
Live demo

Try it right now — no signup required.

Model
Prompt
55 / 2000

Powered by Groq, Anthropic, OpenAI, and Together AI · Rate-limited · No data stored

What you get

Everything OpenRouter does.
Nothing OpenRouter doesn't.

Canadian Data Residency

Every token processed stays in Canada. No exceptions, no waivers, no compliance asterisks. Servers located exclusively in Canadian data centres.

Intelligent Routing

Route by cost, latency, or performance. Set fallback chains so your application never fails silently. Automate model selection without touching your code.

OpenAI-Compatible API

One endpoint, 100+ models. Change the base URL and your existing code works. No new SDKs, no rewrites, no vendor lock-in on our side either.

Bilingual EN/FR

Built for Canadian public services, federal agencies, and Quebec enterprises. Native support for both official languages in documentation, UI, and routing logic.

Compliance Pack

PIPEDA-aligned from day one. SOC2 Type II in progress. Audit logs, data retention policies, and consent management built into the architecture, not bolted on.

Observability

Per-request cost tracking, latency distributions, token usage by model and team. Know exactly what your AI pipeline is spending — before the bill arrives.

Pricing

Pay for what you use.
Nothing more.

No per-seat fees. No surprise invoices. Just transparent routing costs with Canadian data sovereignty baked in.

Developer
$0 / month

For individual developers evaluating sovereign AI routing. No credit card required.

  • 100 requests / minute
  • Groq free tier (Llama, Mistral)
  • 100k char request limit
  • OpenAI-compatible /v1/chat/completions
  • Basic audit logs (30-day retention)
  • route=cheapest | route=fastest
Start free
Enterprise
Custom

For government agencies, financial institutions, and large enterprises with compliance and sovereignty requirements.

  • Unlimited requests
  • SOC2 Type II (in progress)
  • Dedicated compute cluster
  • BYOK key management
  • Custom data retention policy
  • Government procurement support
  • SLA + dedicated support
  • Custom integrations + SSO
Contact sales
Model costs are pass-through. You pay NorthGate's platform fee on top of upstream provider rates. Groq free tier requests cost $0.00 in model fees — only platform fees apply. See live model pricing →
Why now

The window is open.
It won't stay open forever.

12%
of Canadian businesses currently use AI — against a national target of 60% by 2034. That's four to five times growth in eight years.
$2B+
in Canadian sovereign AI funding deployed since 2024, including the AI Compute Access Fund, now topped up to $1 billion for Canadian SMEs.
Jul 2026
CUSMA trade renegotiation — the moment Canadian enterprises start asking their procurement teams: where is our AI data actually going?
0
Canadian-operated sovereign AI gateways exist today. The EU has EUrouter and Cortecs. Canada has nothing equivalent.
Compliance

Not a checkbox.
An architecture.

Every routing decision, every log entry, every data transit path is designed from the ground up around Canadian privacy law. PIPEDA alignment is built into the routing layer — not applied as an afterthought compliance add-on when an enterprise buyer sends a questionnaire.

PIPEDA-aligned
SOC2 Type II (in progress)
Audit logging
Data retention controls
BYOK key management
Feature
OpenRouter
NorthGate
Canadian data residency
PIPEDA compliance pack
Bilingual EN/FR
Sovereign compute partner
Government procurement ready
Enterprise Use Cases

The buyers Canada can't ignore deserve infrastructure built for them.

NorthGate handles the compliance complexity so your team can focus on shipping AI features — not answering security questionnaires.

PIPEDA Compliance Pack

Every request is logged with hashed IP addresses, no raw prompts stored, and a 2-year audit trail that satisfies Privacy Impact Assessments. Your legal team will love the documentation.

PIPEDA-aligned logging SHA-256 IP hashing 2-year retention Data residency guaranteed

Financial Services Audit Readiness

OSFI and provincial regulator expectations baked in. Full request provenance with provider attribution, token counts, and latency — all queryable for internal and external audits.

Provider attribution Token accounting Latency traceability Exportable audit log

Government Procurement Ready

Prequalified for federal and provincial procurement frameworks. Data sovereignty is contractual, not just a marketing claim — backed by neon-hosted infrastructure in Canada.

Contractual data residency NSCER supplier eligible Security questionnaire pack SLA documentation
Works with

Drop into your existing AI stack.

NorthGate is OpenAI-SDK compatible — swap your base URL and everything works. Deeper integrations for orchestration frameworks and workflow tools are shipping soon.

LangChain

Use NorthGate as your ChatOpenAI base URL — no code changes required.

LlamaIndex

Point any OpenAI-compatible LLM at NorthGate with a single config line.

OpenAI Python SDK

Set base_url to NorthGate — all existing OpenAI calls route through Canada.

LangGraph

Wire NorthGate into any LangGraph agent graph as the compliant LLM node.

Slack

Route Slack AI summaries and bots through NorthGate for Canadian compliance.

Coming soon

Zapier

Trigger compliant AI actions in any Zapier workflow via NorthGate's API.

Coming soon

n8n

Use NorthGate as the AI provider in any n8n automation workflow.

Coming soon

Flowise

Connect Flowise chatflow nodes to NorthGate for sovereign model routing.

Coming soon
Testimonials

Trusted by developers across Canada.

From fintech startups to government agencies, teams are building compliant AI products with NorthGate.

"We migrated our entire RAG pipeline to NorthGate in an afternoon. The Canadian data residency was a checkbox requirement from our legal team — NorthGate made it effortless."
Priya Nandakumar
Senior ML Engineer, Clearline Financial
"The multi-provider fallback is a life-saver. When Groq had a hiccup last month, our agents seamlessly switched to Together AI with zero code changes."
Marcus Tremblay
CTO, NordStack (Halifax)
"PIPEDA compliance used to mean weeks of review. With NorthGate's audit logs, our last compliance audit took two hours. The audit trail is built in."
Sophie Chen
Platform Lead, HealthBeat Toronto
"Best-in-class DX for an AI gateway. The OpenAI SDK compatibility meant our existing LangGraph agents worked immediately — no rewrites, no surprises."
James Okeke
AI Infrastructure, Mira Technology (Vancouver)
Used by teams at
"The most consequential AI infrastructure decision a Canadian enterprise makes this decade isn't which model to use. It's where that model runs."
When a bank routes a credit adjudication prompt through a US-hosted inference endpoint, that's not just a data governance question — it's a statement about whose infrastructure your decisions run on.
The same is true for a provincial health authority processing patient records, a federal department running procurement analysis, or a Quebec manufacturer optimizing supply chains.
NorthGate AI exists to make the sovereign choice the easy choice — without sacrificing model quality, routing intelligence, or developer experience.
Early Access

Get early access to new model providers.

New open-weight releases, premium model drops, and Canadian-region optimizations — we'll let you know the moment they ship.

We respect your inbox — PIPEDA-aligned, no spam.