Stack
We don’t change tools every quarter. This is the stable stack we design, build and operate what we deliver with, and the criterion we choose it by.
The criterion: reliability, security and staying in control
Section titled “The criterion: reliability, security and staying in control”We choose every piece of the stack for its reliability, its security and how well it fits the problem in front of us, favoring open standards and solutions that keep the technical control in-house. We avoid locking ourselves into a single vendor whenever we can, and when it’s worth paying for a tool because it’s the best fit for the case, we tell you why we picked it.
Automation
Section titled “Automation”Our automation platform is the base we build most flows on: a visual, open source builder that connects hundreds of tools (email, CRM, spreadsheets, messaging, storage, ecommerce, accounting) and lets us add custom logic with code nodes when a connector doesn’t cover the case. We deploy it self-hosted on European infrastructure, so the flows and the data passing through them don’t depend on a third party outside the EU. It’s also the base of our own internal automations, so we test it first on our own business before putting it on yours.
- Open source at its core: auditable, no vendor lock-in.
- Self-hostable in EU territory, so sensitive data doesn’t leave where it should stay.
- Connects with practically any API or database you already use.
- Custom logic with code nodes, without having to rewrite the whole platform for a specific case.
Backend and data
Section titled “Backend and data”- PostgreSQL as the main database.
- Supabase for Auth, Realtime and managed Postgres when the project calls for it.
- Docker to deploy the automation platform and our own services in a reproducible way.
Frontend
Section titled “Frontend”- React 18 + TypeScript + Vite for the product side.
- Tailwind CSS + shadcn/ui for consistent, fast-to-maintain interfaces.
Artificial intelligence
Section titled “Artificial intelligence”We work with an abstraction layer over multiple model providers, exposed with stable aliases (Prelude, Sonata, Symphony, Concerto) instead of the real provider’s name. This protects you as a client: if a provider raises prices, degrades a model or discontinues it, we swap it for an equivalent one without you having to change anything on your side. For the most demanding case, we also offer local models that run on your own infrastructure, so the data never leaves your perimeter.
Infrastructure and operation
Section titled “Infrastructure and operation”- VPS in EU territory for the services we operate ourselves.
- Cloudflare for CDN, Tunnel and website deployment (Pages).
- Public status with historical uptime, so you can check availability without asking us.
- Telegram for internal operational alerts when something needs human attention.
- Resend for transactional email, with DKIM, SPF and DMARC configured on our own domain.
The detail of how we protect what runs on this stack (encryption, per-client isolation, backups, monitoring) is in Security & data and, with full technical detail, in the Trust Center.