Custom
Speech-to-Text.
High-performance speech-to-text systems engineered for real results, from prototype to production deployment.
Execution Model
Built in Partnership
Robust deployment without generic compromises.
We collaborate closely with your domain experts to build speech-to-text solutions that address your specific operational hurdles.
Production History
Featured Projects
AI systems built in partnership across diverse industry domains.

Murmel
ProductA frontier Dutch speech-to-text AI model, trained on thousands of hours of spoken Dutch. Fully hosted in Europe with no dependency on external parties.
Key Features
- Optimised for Dutch speech
- Fully hosted in the EU
- No dependency on external parties

FirstRing
IndustryTogether with crisis communication experts, we built an AI-powered platform that generates response plans and messages within seconds. Engineered for safety-critical environments.
Core Components
- Rapid Response Generation
- Model agnostic with fallback systems
- Optimised for the crisis communication domain

Beleidsradar.nl
VentureAn AI-powered intelligence platform monitoring millions of policy documents. Features advanced vector search and LLM-driven synthesis for legislative monitoring in The Hague and Brussels.
Technical Stack
- High quality debate transcriptions
- Agentic system for alerting users of relevant developments

NPO Radio AI
Client ProjectIn partnership with NPO's editorial team, we built AI systems that unlock archived audio in seconds and accelerate radio production workflows.
Technical Stack
- Making large quantities of audio searchable
- Editorial workflow with AI assistants
- Large scale search engines

Applied Research
ResearchInvestigating the infiltration of synthetic media in global search results. Our technical audit for De Volkskrant exposed the critical challenges to information integrity in an AI-saturated world.
Research Impact
- Societal impact of AI
- Large scale data collection
- Automatic detection of generative AI

ChatZaanstad
IndustryBuilt in partnership with Gemeente Zaanstad: a secure, model agnostic AI chat interface that enables local governments to manage their own AI models hosted entirely within the Netherlands or the EU.
Key Features
- Model Agnostic (Llama, Mistral, GPT)
- Prompt Library for shared expertise
- Custom styling based on organization

Computer Vision
ProductTogether with our clients, we engineer production-grade computer vision systems for organizations that cannot afford black boxes, vendor lock-in, or fragile demos.
From data pipelines and model training to deployment on sovereign infrastructure, we design systems that remain auditable, scalable, and under your control, even in high-stakes operational environments.
Core Focus
- End-to-End CV System DesignData ingestion, annotation strategies, model training, evaluation, and deployment.
- Sovereign & On-Prem DeploymentFull control over models, data, and runtime. No third-party dependency.
- High-Reliability Visual IntelligenceBuilt for real-world variability, not lab conditions.
Process
How Does It Work?
From first idea to production: a clear and transparent process in 8 steps.
Share Your Vision
Send us a message with information about what you want to build. We'll review your idea and get back to you quickly.
Discovery Meeting
We schedule a meeting to discuss the possibilities. Where should it be hosted? How does it work? We map out the technical landscape together.
Quote & Mockups
Based on the meeting, you receive a detailed quote and several mockups showing how the solution could work.
Review & Refine
We discuss the project timeline and adjust anything you'd like to see different. Your input shapes the final direction.
Development Kickoff
After you agree to get started, our team begins development. We set up the infrastructure and start building.
Iterative Development
During development we keep you up to date with progress. You can test the product and give feedback in real time.
Launch & Delivery
When everything is exactly how you want it, we either hand over the project to your technical team, or we can host and maintain it for you at a great price.
Grow & Expand
Want to improve or expand? That's always possible. We're here for the long run to help you scale your AI capabilities.
Ready to start building? Our process is designed to be transparent, collaborative, and always centered on your goals.
Get StartedFAQ
Frequently Asked Questions
A functional MVP typically takes 3–6 weeks. Complex integrations may take longer, depending on your existing infrastructure.
Yes. We specialize in sovereign deployments on EU-based clouds or complete on-premises hardware without third-party dependencies.
Absolutely. We ensure your engineering team is fully equipped to maintain and scale the system independently after launch.
Ready to move AI from pilot to production?
A 30-minute intake. We map your use case, sketch the architecture, and show what's possible. No sales pitch. You leave with a concrete plan.
Schedule a free intake (30 min)