INTELLIGENCE AGENCY FOR AI INFRASTRUCTURE
A Canadian Intelligence Firm
Vanwebdev LTD is a Canadian intelligence agency. We build, defend, and brief Canadian AI developers on production AI infrastructure aligned with PIPEDA Compliance.
THREE TRUNKS
Build. Defend. Brief.
One discipline, three trunks. Each pillar links to its own page with deeper methodology and pricing.
HEADLINE SERVICES
Four named engagements operators run before they ship
Three more services sit under /pricing/. These four are the headline engagements you can book in a single phone call.
Operations Intelligence
AI workflow audit mapping bottlenecks and shipping a 30-60-90 rollout brief.
$3500 fixed
Threat Brief
Monthly threat intelligence brief mapped to PIPEDA, Bill C-27, and ITSG-33.
$350 per month
Sovereign AI Defense
Runtime monitoring retainer. Prompt injection, jailbreak, exfiltration alerts.
$1500 per month
FLAGSHIP PRODUCTS
Four product engagements ship under named SKUs
Each product is a scoped engagement, not a license. Pricing reflects the work, not the seat-count.
Sovereign AI Box
H100 GPU bundled with an open-weight LLM and an agent runtime, installed on your hardware.
From $42,000
Open Claw Pro
Production-grade Open Claw tier with HA failover, observability, secrets, and on-call webhooks.
From $18,000
Agent Deployment
We build the agent, ship it to production, wire observability, and hand over the runbook.
From $12,000
PivotToAI Seat
Ninety days of 1:1 coaching, three shipped milestone projects, and an evaluation framework.
$8,000 cohort
CONTENT HUBS
Five hubs operators check on a monthly cadence
Free to read. No gate. No upsell baked into the dispatch.
WHY THIS POSITIONING
Canadian AI intelligence agency, not Canadian AI agency
The Canadian AI intelligence agency framing carries weight because it sets the contract. An agency ships work; an intelligence agency ships work plus a written assessment of what an adversary would do next. Specifically, every Vanwebdev engagement ends in operating infrastructure plus a named threat-model document.
Furthermore, the intelligence-agency posture means we publish before we sell. Read the latest Threat Brief on the brief hub, browse the case-studies hub, or download an artifact from the library hub. Additionally, the reading dispatch ships every Tuesday with no upsell embedded.
Consequently, operators who book the $250 Intelligence Audit are not buying a sales call. They are buying a 60-minute working session that lands a risk map and a 90-day plan, with the option to take that plan to another firm if Vanwebdev is not the right fit.
AGENTIC SYSTEMS
If you are scoping agent work, start here
Agentic AI infrastructure is the named-engagement track for operators who already know they need an agent in production. Specifically, the agentic-systems pillar covers tool-use orchestration, persistence layers, evaluation loops, and on-call observability for agent fleets running across multiple Canadian regions.
Notably, every agentic engagement ships a build manifest plus a runbook your operations team owns. Likewise, every deployment includes a reproducible eval suite mapped to the operator’s named risk surface, not a vendor-supplied benchmark.
RESEARCH
Reproducible Canadian AI research
All methodologies reproducible. All datasets versioned. All artefacts free to download.
From the research pillar
JURISDICTION BY DESIGN
Why a Canadian AI intelligence agency matters
PIPEDA, Bill C-27 AIDA, and ITSG-33 are not compliance theatre at Vanwebdev. Specifically, each engagement maps the operator’s data flows against PIPEDA principles three through ten and against the Protected B handling baseline where applicable. Additionally, every build manifest names the Canadian region of inference and the operator-owned hardware boundary.
Notably, Vanwebdev does not hold operator model weights on a US hyperscaler. Consequently, the build trunk delivers inference layers on Canadian regions and on operator-owned hardware. Furthermore, the about page documents the team and the methodology at named-bio depth.
Importantly, the federal Bill C-27 Artificial Intelligence and Data Act citation lives at parl.ca and we re-curl it every wave to confirm the URL holds. Likewise, the OPC PIPEDA citation lives at priv.gc.ca and anchors the privacy posture. See Bill C-27 at parl.ca and OPC PIPEDA.
CASE STUDY SPOTLIGHT
OnSiteSafety: a $5K class became a productized AI stack
Specifically, the OnSiteSafety engagement turned a Vancouver-region in-person safety class into a productized AI stack with named regional inference, an operator-owned data layer, and a 90-day plan that shipped on schedule. Notably, the case study demonstrates how a Canadian AI intelligence agency reframes the engagement contract: methodology end to end, named artifacts, and reproducible eval suites the operations team owns post hand-off.
FROM THE LIBRARY
Three artifacts to read this week
AI Risk Map for Canadian Operators
A reusable risk-map template mapped to PIPEDA principles and ITSG-33 controls.
Operator Glossary of Production AI Terms
Plain-language definitions of the 80 production AI terms operators actually use.
Monthly Canadian Threat Brief
The latest threat brief covering CSE advisories, Bill C-27 movement, and named attacks.
START HERE
Book the audit. Read the brief. Decide later.
Specifically, the Intelligence Audit is a $250 sixty-minute working session that lands a risk map and a 90-day plan. Notably, a Canadian AI intelligence agency pays back the engagement fee inside the first decision; there is no commitment to a follow-on retainer.