We Collected Every Job in Sweden. Then We Regretted It.
We aggregated 14,000+ assignments and built a bigger haystack. Here's how we fixed it with AI agents.
Three years ago, we set out to solve a simple problem: fragmentation. Consultant assignments were scattered across hundreds of broker sites, portals, and locked PDFs.
So we built aggregators. We collected 14,000+ live contracts—practically the entire Nordic market in one place.
But we made a mistake.
We thought "more data" was the solution. It wasn't. By centralizing everything, we just built a bigger haystack. Finding a specific niche role—like "Rust developer in Malmö willing to work onsite but only 3 days a week"—was still a nightmare of keywords, filters, and manual scrolling.
We had the data. Finding the needle was still manual labor.
From database to agent
We realized that no amount of UI filters could solve this complexity. Checkboxes and dropdowns can't filter for context.
But an LLM can.
We didn't just want better search. We wanted to give every freelancer their own 24/7 assistant. Something that reads every new assignment, understands your specific constraints, and tells you only when there's a match worth your time.
How it works
We built an MCP server that exposes our entire live database directly to AI assistants like Claude.
Before: You visit the site daily, type "Python," scan 50 listings.
After: You tell your assistant: "Monitor for senior Python gigs in Stockholm. Ignore banking clients. Ignore maintenance roles. Alert me if something looks interesting."
It turns job hunting from a pull process (you searching) into a push process (the agent notifying you).
See it in action
Ask your AI assistant to find you assignments. No filters, no keywords—just describe what you're looking for in plain language.
"Find me remote React contracts in Sweden, at least 900 SEK/hour, no banking clients."
The assistant searches our entire database and returns only what matches.