Today, we're excited to announce that Weflayr has raised $2.5 million in pre-seed funding, led by 13books Capital, with participation from Founders Future, Kima Ventures, Plug and Play, Station F Ventures, Roxanne Varza (General Catalyst scout), and a group of incredible angel investors.
This is a major milestone for us, and just the beginning of our journey to give AI companies full control over their inference strategy.
AI models are evolving fast. The infrastructure isn't keeping up.
Hundreds of AI models now compete on quality, speed, and cost. New releases are constantly reshaping what's possible.
Yet most AI companies still rely on just a handful of models to power the majority of their workloads.
Why? Because quality is non-negotiable.
Different models excel at different tasks, but knowing which model performs best for your specific use case remains a challenge.
Traditional AI gateways can route requests between models. But without understanding how those models actually perform on your own workloads, routing decisions remain largely blind.
We believe AI companies shouldn't have to choose between quality and cost.
Introducing a smarter way to route AI inference
We're building Weflayr to give AI companies full control over how they evaluate, select, and route AI models.
Instead of relying on generic benchmarks or black-box routing, Weflayr learns from your own workloads.
Here's how it works:
- Benchmark on your own data. Evaluate different AI models against your real production workloads, based on your quality and cost objectives.
- Stay in control. Define your routing strategy and approve the models you trust.
- Route intelligently. Weflayr dynamically matches incoming requests to similar previously benchmarked calls, selecting the best-performing approved model within your defined boundaries.
- Optimize your economics. Connect every inference decision to its business context, with visibility into costs and margins across customers, features, and workflows.
Early results speak for themselves
Our first deployment has demonstrated promising results:
Over 55% lower inference costs while maintaining equivalent quality.
15 percentage points higher relative quality at equivalent cost compared with a frontier model.
Our ambition goes beyond building another AI gateway. We want to create the intelligence layer that enables AI companies to continuously optimize performance, cost, and margins.
Built by founders who know critical infrastructure
Weflayr was founded by Romain Mazeries, former CEO of Mangopay, and Noé Achache, former Head of AI Engineering at Theodo Data & AI.
We've spent our careers building and scaling complex infrastructure, from payments to production AI systems.
We've seen how intelligent routing transformed the payments industry. We believe AI inference is following a similar path, at a much greater speed.
And we're excited to help define what comes next.
Backed by investors who share our ambition
We're incredibly grateful to 13books Capital, Founders Future, Kima Ventures, Plug and Play, Station F Ventures, Roxanne Varza, and all our angel investors for believing in our vision.
“AI inference is rapidly becoming a critical infrastructure layer for the next generation of software companies. As the market evolves and consolidates, we believe there is a massive opportunity for an independent intelligence layer that gives AI companies full control over quality, performance, and economics.”
Mike McFadgen, Partner at 13books Capital
This funding will help us expand our engineering and AI research teams, accelerate product development, and bring Weflayr to more AI companies across Europe and the United States.
What's next?
We're already onboarding our first customers and preparing for broader product availability in the coming weeks.
Our mission is simple: make AI inference smarter, more efficient, and fully under the control of AI builders.
And we're only getting started.
Join us on this journey.
