Civo has unveiled the relaxAI API, designed to deliver cost-effective generative AI capabilities to UK organizations while ensuring data control and regulatory compliance. The API is OpenAI compatible and offers developers sophisticated tools like Llama 4 Maverick 17B-128E, DeepSeek R1 0528, and Mistral-7b-embedding. These tools smoothly integrate into applications, enabling functionalities like code generation, sentiment analysis, and advanced reasoning.
The standout feature of relaxAI is its sovereign infrastructure, hosted entirely in UK data centers, ensuring all AI processing occurs within the UK's legal jurisdiction. This initiative trails the rising skepticism among UK IT leaders regarding reliance on US AI platforms. A survey conducted by Civo indicated that 67% of over 1,000 respondents expressed concerns over data ownership, preferring AI services that guarantee data control.
The relaxAI API addresses these apprehensions through UK-exclusive hosting, comprehensive security measures like end-to-end encryption, multi-factor authentication, and adherence to standards like GDPR, SOC 2, CyberEssentials Plus, and ISO-27001. Users benefit from flexible deployment options, including cloud-based, on-premises, or hybrid implementations.
Performance and scalability are integral to the API's design. It is among the few UK-hosted solutions supporting Llama 4 Maverick and similar models while integrating seamlessly with over 20 industry-recognized frameworks like LangChain, LlamaIndex, GPT4All, and Rasa.
Josh Mesout, Civo's Chief Innovation Officer, emphasized the need for an AI platform combining easy integration with assured data ownership. He affirmed, "With relaxAI, we’re offering developers and enterprises the best of both worlds: world-class AI capabilities with the guarantees of domestic infrastructure and compliance. Our goal is to make sovereign AI not just secure, but simple.."
The relaxAI API, currently available for relaxAI Free and Plus users, offers UK-based deployments competitively priced, starting at £0.10 per million input tokens and £0.60 per million output tokens—a fraction of the cost compared to AWS, Google, and OpenAI models.
