Codebase Memory · Context Intelligence
Modern AI coding tools are powerful but amnesiac. They don't know why the code was built the way it was. VAEKTOR connects your entire decision chain — RFC, PR, ADR, code, API logs — and exposes it via MCP so every AI tool has precise context.
// the problem
GitHub sees code. Jira sees tickets. Notion sees documents. None of them talk to each other — and AI tools only see whatever ends up in the current context window. VAEKTOR connects the entire chain.
← VAEKTOR embeds and indexes all of this. Your AI tools query it via MCP. →
// what VAEKTOR does
One system connecting your code, decisions, and quality assurance. Accessible by humans and AI alike.
Ask "how does authentication work?" and get ranked results with code context. Natural language across all your repositories — not grep.
Store every architectural decision with full context — the problem, chosen solution, and rejected alternatives. Survives developer turnover. Queryable years later.
Joins production API logs with your codebase. Know which endpoints are actually used, which are dead weight, and where real traffic flows — not just what the code says.
// mcp integration
Without context, AI tools read entire files, grep blindly, and ask for clarification — burning tokens on every request. VAEKTOR connects via MCP so your tools get exactly what they need, nothing more.
// deployment
Self-hosted is not a compromise — it's premium for teams with compliance requirements. All three models use the same codebase.
Hosted on Hetzner Frankfurt — EU, GDPR-compliant. Connect and go. No infrastructure to manage.
Solo & Team plansDocker Compose image on your own infrastructure. Your code never leaves your network. Full data control.
Business planSame Docker image, your infrastructure. Includes SLA, DPA agreement, dedicated onboarding, and annual contract.
Enterprise plan// roi calculator
Adjust the sliders to match your team's usage and see what VAEKTOR's ~70% token reduction means for your bill.
// your estimate
Based on ~70% token reduction. Actual savings vary by usage pattern.
// pricing
One consultant day costs 8000–12000 DKK. VAEKTOR Business pays for itself the first time it saves your team a day of repeated work.
Self-hosted Docker image · Community support · Core only · Unlimited repos
1 developer · 5 repos · Cloud-hosted · RFC + PR modules
10 developers · 30 repos · Cloud-hosted · All Team modules
Unlimited developers · Cloud + self-hosted · Compliance · SLA 99.9%
// built for
AI answers that know your actual codebase, not a generic one. Ask in plain English, get precise context.
Decision history that survives developer turnover. RFC decisions, rejected alternatives, PR rationale — all queryable.
Measurable savings on AI API costs and consultant days. Architecture drift detection and compliance evidence built in.
GDPR-native, EU-hosted, full data control. Self-hosted option means your code never leaves your network.
// vs. alternatives
| Mem0 / Zep | VAEKTOR | |
|---|---|---|
| Data location | US cloud | EU-hosted or self-hosted |
| What it remembers | Conversation history only | RFC → PR → Code → API logs |
| API log integration | ✗ | ✓ Pulse module |
| Self-hosted | Limited / paid | First-class, Docker Compose |
| GDPR / DPA | On request | Included on Business+ |
| MCP support | Limited | Native, all tools |
// the name
pronounced /ˈvɛktər/ · AE = Æ, the Danish vowel
They are homophones — pronounced identically. The Danish letter Æ replaces the English E while keeping the exact same sound.
In mathematics and physics, a vector is a directed quantity — something with both magnitude and direction. That is what VAEKTOR gives your AI tools: not just raw code, but context that points in the right direction.
Your AI tools are expensive when they're guessing. VAEKTOR fixes that.