Weflayr Raises $2.5M Pre-Seed Led by 13books Capital for AI Inference Routing

Weflayr Closes $2.5M Pre-Seed

Paris-based AI infrastructure startup Weflayr has raised a $2.5 million pre-seed round led by 13books Capital, per funding data dated October 22, 2026. The round brings together a mix of venture funds, an accelerator, and an angel, giving the company capital to develop its AI gateway for inference routing.

What Weflayr Does

Weflayr builds an AI gateway that routes each inference request to the model best suited for it. The workflow is designed around how AI teams actually operate: teams benchmark candidate models against their own production workloads under quality and cost objectives, then approve the models they trust. The gateway then dynamically matches incoming requests to the best-performing approved model, factoring in business context such as the customer and the margin attached to a given request.

In plain terms, Weflayr is betting that model selection should be continuous and workload-specific rather than a one-time architectural decision. As the number of available models grows, routing becomes a cost and quality lever rather than a static configuration.

Funding Details

• Round type: Pre-Seed
• Amount: $2,500,000
• Currency: USD
• Date: October 22, 2026
• Lead investor: 13books Capital
• Investors: Plug and Play, Founders Future, Station F, Kima Ventures, 13books Capital, Roxanne Varza

Founding Team

Weflayr was founded by Romain Mazeries, former CEO of Mangopay, and Noe Achache, former head of AI engineering at Theodo Data & AI. The pairing combines payments-scale operating experience with hands-on AI engineering leadership, a combination that maps directly onto the company's pitch around cost and margin-aware routing.

Early Results and Program Affiliation

The company reports that its first deployment cut inference costs by more than 55% at equivalent quality. That figure, if it holds across broader workloads, speaks to the core commercial argument for routing infrastructure: inference spend is now a material line item for AI-heavy businesses, and even modest efficiency gains compound at scale.

Weflayr is also part of Station F's F/AI program, and Station F appears among the round's investors alongside Kima Ventures, Plug and Play, Founders Future, and angel Roxanne Varza.

Why Inference Routing Matters

AI teams increasingly run multiple models rather than committing to a single provider. That creates a coordination problem: which model handles which request, at what cost, and at what quality threshold? Weflayr's answer is an approval-based gateway where teams retain control over which models are eligible, while the routing layer handles the moment-to-moment matching.

The inclusion of business context, such as the customer and margin attached to a request, is a notable design choice. It suggests the gateway is aimed at production businesses with heterogeneous workloads rather than at pure research settings.

Acquisition Angle

AI infrastructure is an active area for strategic buyers, particularly among cloud providers, model providers, and developer-tooling platforms that want to sit between applications and models. A gateway that owns routing logic and accumulates benchmarking data on model performance could be an attractive asset, since it sits at a point in the stack where switching costs can build.

At pre-seed, Weflayr is early, and the reported 55% cost reduction comes from a first deployment rather than a broad customer base. Acquirers will want to see repeatable results across diverse workloads and evidence that the approval workflow becomes embedded in how teams operate.

What's Next

The $2.5 million round gives Weflayr room to expand its gateway and bring on more AI teams. The company has not disclosed specific use-of-funds plans. With a founding team drawn from payments and AI engineering, and backing from 13books Capital alongside Station F and Kima Ventures, Weflayr enters the next phase with both operating credibility and ecosystem access.