Kilo Partners with Inceptron for High-Performance EU Inference

Access fast and secure open-weight models in the Kilo Gateway

We are thrilled to announce a strategic partnership with Inceptron to bring top-tier, EU-hosted inference directly to the Kilo platform.

Whether you’re building complex cloud agents or generating code in your IDE, this integration ensures you have access to powerful open-weight models with strict adherence to European data privacy standards.

Introducing Inceptron

Headquartered in Sweden, Inceptron is designed for teams that prioritize both speed and data sovereignty. They’re here to help you run open-source or fine-tuned models on infrastructure purpose-built for production, and they provide sovereign European AI infrastructure as an alternative to US-based cloud providers. Users can explicitly select and lock in data residency regions (such as optional EU-only processing) to maintain compliance and security.

By partnering with Inceptron, Kilo is expanding its European footprint, allowing you to route your prompts through highly optimized infrastructure built from the ground up for AI workloads.

Our partnership with Kilo represents a major milestone in providing developers with high-performance inference that never compromises on European data sovereignty or GDPR compliance. Together, we are ensuring that privacy-conscious teams can leverage the world’s most capable open-weight models in the Kilo Gateway. We use Kilo ourselves for agentic engineering and have found it to be the best solution on the market!

– Nima Karimi, Inceptron

Access Top Open-Weight Models Hosted in the EU

Through this partnership, Kilo users instantly gain access to some of the most capable open-weight models available today, all hosted securely by Inceptron and optimized for both speed and security.

Currently supported models include:

• MoonshotAI: Kimi K2.6

• Zai: GLM 5.1

• MiniMax: MiniMax M2.5

These top open-weight models are daily drivers for many developers. As AI costs continue to rise, we’re seeing everyone from independent devs to large enterprises take advantage of these cost-efficient models. Kimi K2.6, a multimodal model designed for long-horizon coding, has been one of the most popular models on both Kilo and Inceptron since it launched.

We know that no two engineering teams operate exactly alike, which is why we offer two distinct ways to access Inceptron’s hosted models.

Kilo Gateway (Recommended)

For the absolute smoothest experience, you can access Inceptron directly through the Kilo Gateway. This route is popular with organizations that demand model freedom and flexibility without sacrificing top speed or performance. The Gateway handles all the routing and load balancing behind the scenes, letting you switch between models effortlessly. As seen in the platform dashboard, inference is highly cost-effective, with models like MiniMax M2.5 running as low as $0.15 per 1M input tokens.

Bring Your Own Key (BYOK)

If you already have an established relationship and custom billing with Inceptron, you can easily use our BYOK setup. Simply drop your API key into the Kilo dashboard. Just remember to use the specific inceptron-byok/ prefix for your model selection so your client knows exactly where to route the request.

Built for ISO and GDPR Compliance

In the era of enterprise AI, data privacy, security, and local compliance are paramount considerations when buyers are choosing where to host workloads. Regulations like the General Data Protection Regulation (GDPR) are technology-neutral; any AI system that collects, trains on, or processes the personal data of EU residents must comply with its strict privacy and data protection principles.

Because Inceptron’s infrastructure is strictly contained within the European Union, routing your inference through their servers helps your organization maintain strict compliance with GDPR. Furthermore, their infrastructure aligns with rigorous ISO standards, giving enterprise teams the peace of mind needed to deploy AI tools in production environments where data residency is heavily audited.

Inceptron pairs perfectly with Kilo for Enterprise, which lets you meet your team where they already work (VS Code, JetBrains, CLI, Cloud, Slack, Code Reviews, and more) along with SSO, audit logs and a dedicated SLA.

Usable Across the Entire Kilo Code Ecosystem

The best part about this integration is that it works wherever you already work.

Once Inceptron is enabled in your Kilo dashboard—whether via the Gateway or BYOK—these models are immediately accessible across the entire Kilo Code suite:

• Kilo CLI: Run quick, secure inferences directly from your terminal.

• Cloud Agents: Power your automated workflows with EU-hosted intelligence.

• IDE Extensions: Generate code natively inside JetBrains and VS Code without your proprietary source code leaving the European Union.

Head over to your Kilo dashboard today to enable Inceptron in your Providers settings. Looking for Kilo for Enterprise? Book a demo and we can dive into the details!