91 Tools, One Server: How We Use MCP to Run Sales, Invoicing, and Inventory From a Single AI Interface
Six modules. Four backend types. Two transports. One binary. Here's how we wired our entire operations stack into a single MCP server.
Hugo MatthaeyBuilding Smarter Context: Search Strategies That Make AI Actually Useful
No single search strategy is good enough. Combining semantic, structured, and fuzzy search hides each method's weaknesses while stacking their strengths.
German Skrylnikov
We Built Our CRM From Scratch. Here's What 64 Versions Taught Us About Outbound
Sales tools are everywhere. Most are built by people who've never sent a cold message. We built our own — 64 releases later, here's what's inside.
Hugo Matthaey
What 30,000 Messy Wine Entries Actually Look Like
We wrote about helping a wine ERP stop losing weekends to data cleanup. This post goes deeper — into the data itself.
Armi Piamonte
Beyond Text: Solving Real-World Document Understanding with Vision LLMs
What if a single multimodal vision model could understand any document thrown at it? Here's what we learned building production-grade image recognition across wine labels, financial documents, and handwritten schedules.
German Skrylnikov
Transcribing Mixed Arabic-English Voice Messages: What We Learned from Real UAE Grocery Orders
We benchmarked Whisper, fine-tuned models, ElevenLabs, and Hamsa against 25 real WhatsApp voice notes — none hit production-grade accuracy.
German Skrylnikov 
How We Helped a Wine ERP Stop Losing Weekends to Data Cleanup
30,000 messy wine entries. Weekends lost to manual fixes. Here's how we turned it around.
Hugo Matthaey
