The AI systems we test, build with and work in every day

The AI field changes every week. We test new models and tools continuously — and only build on what holds up in production. These are the systems and platforms we use to deliver your solutions.

AI models

The language models we build on

We choose the model to fit the task — not out of habit. And we switch easily when new models beat the old ones.

🟢

OpenAI — GPT series

Our workhorse for conversation, reasoning, TTS (voice) and API integrations. GPT-4o and the latest models in production with us.

🟠

Anthropic — Claude

Top class for coding, long contexts and complex analysis. We use Claude for AI-assisted development and agent solutions.

🔵

Google — Gemini

Strong at multimodal tasks: images, video, documents and audio. Good price/performance at scale.

⚫

DeepSeek & open source

Llama, Mistral, DeepSeek — when cost or data residency is decisive. Can run on your own server.

Voice & telephony

Voice AI and telephony — our speciality

We have built an entire voice platform (Agento), so we know every layer of voice architecture.

🎙️

Deepgram

Real-time speech-to-text (STT) — with Danish support and low latency. The foundation of our voice agents.

🗣️

ElevenLabs

Ultra-natural voices for text-to-speech (TTS). We have built 6 Danish elite voices for our platforms.

📞

Telnyx

Telephony infrastructure for AI calls: WebSocket, SIP and numbers. Our preferred provider for voice agents.

📱

Twilio

Alternative telephony stack with SMS, voice and WhatsApp. Used when the client is already on Twilio.

Development & backend

The technical stack we build in

⚛️

React & Next.js

Our frontend foundation: lightning-fast, SEO-friendly apps and websites with server-side rendering.

🟢

Node.js & TypeScript

Robust backends, APIs and real-time systems with type safety — built to last.

🐍

Python

AI pipelines, data analysis, automation and LLM integration. Python is the native language of AI.

🗄️

Supabase & PostgreSQL

Open-source backend: database, auth, storage and real-time — with full control over your data.

🔀

n8n

Self-hosted workflow automation — with AI nodes and full control over the data flow.

🔗

Make & Zapier

Fast no-code integrations between 5,000+ apps — for when time matters more than control.

Infrastructure & hosting

Where your systems run

☁️

OVHcloud

Our preferred EU server provider. Your data in Germany — GDPR-safe, scalable and affordable.

🇩🇪

Hetzner

German cloud infrastructure known for price and performance. Used for webshops and high-traffic sites.

🐳

Docker

Containerisation that makes deployment and scaling reproducible — wherever the system runs.

🔐

Nginx & Caddy

Reverse proxy, automatic SSL renewal, caching and load balancing — the server layer that never fails.

⚡

Vercel

Edge hosting for Next.js apps with a global CDN — for frontends that need to be extremely fast.

🛡️

Cloudflare

CDN, DDoS protection and DNS in front of your systems — world-class security.

Tools

Tools we use every day

🖥️

VS Code & Cursor

AI-powered code editors — we code with AI assistance all day, every day.

👥

GitHub & CI/CD

Version control, code review and automated deploy pipelines — everything goes through git.

📊

Google Analytics & Search Console

Data on traffic, search and conversions — the foundation for every SEO decision.

📈

Google Ads & tagging

Advanced ad optimisation and conversion tracking — we are Advanced-certified.

🎨

Figma

Design and prototyping tool — from sketch to clickable prototype before any code is written.

💬

Slack & MS Teams

Communication and integrations — and often where your automations send notifications.

Why we test AI systems continuously

The AI landscape is changing faster than any other field of technology. A model that was best in class three months ago may be overtaken today — and prices change constantly. That is why we spend time every week testing new models, tools and workflows.

That gives you three concrete advantages: 1) You always get the best technology for your task — not what we learned two years ago. 2) We can switch model or vendor without rewriting your solution, because we build on an abstraction layer. 3) You get honest recommendations — even when the answer is “you don’t need AI for this”.

Our approach to new AI tools

  • Tested in production: We run our own platforms on new technology before we recommend it to clients.
  • Proven value required: A tool has to prove its value in time, cost and quality — hype is not enough.
  • Data residency: Your data stays in the EU unless otherwise agreed. Always.
  • Documentation: We document what we build with, so you are not dependent on a single vendor.
In short: You don’t need to keep up with AI developments. We do — so you can focus on your business.

Unsure which AI tools to use?

We guide you through the jungle — and only recommend what genuinely creates value for your business.