PREM Confidential API is an AI inference platform operated by PREM SA. Its privacy policy lists Nebius B.V. as a zero-data-retention inference provider processing prompts and completions in the EU/UK. The register specifies transient processing for inference only, zero retention under a data processing agreement and no training use.
PREM’s documentation describes its zero-data-retention mode as an OpenAI-compatible gateway that sends requests to partners operating inference capacity. It distinguishes this mode from its confidential inference mode using trusted execution environments. The documentation does not name individual partners.
PREM Confidential API’s privacy policy names PREM SA as its operator and lists Nebius B.V. as an inference subprocessor. It specifies transient processing of prompts and completions in the EU/UK, zero retention under a DPA and no training use.
PREM Confidential API privacy policyListed 20 Oct 2026 · PREM · Web pagePREM Confidential API’s privacy policy names PREM SA as its operator and lists Nebius B.V. as an inference subprocessor. It specifies transient processing of prompts and completions in the EU/UK, zero retention under a DPA and no training use.
PREM Confidential API is an AI inference platform operated by PREM SA in Lugano, Switzerland. It gives developers and businesses access to language models through an API and dashboard. Its confidential inference mode encrypts requests on the client and processes them inside hardware-isolated trusted execution environments. A separate OpenAI-compatible zero-data-retention mode routes requests to inference partners under contractual retention restrictions. Developers can integrate model responses into applications while choosing between these different privacy guarantees.
Prem AI launches CyberScan for continuous code security audits
Prem AI launched CyberScan in beta for continuous security reviews, initially targeting Bitcoin infrastructure. It supports resumable repository scans, prioritised findings, and integration with GitHub, APIs and MCP. ArkLabs and Breez tested it against production code before launch.
An AI inference platform operated by Swiss company PREM SA. · Infrastructure
How they use Nebius Token Factory
PREM Confidential API is an AI inference platform operated by PREM SA. Its privacy policy lists Nebius B.V. as a zero-data-retention inference provider processing prompts and completions in the EU/UK. The register specifies transient processing for inference only, zero retention under a data processing agreement and no training use.
PREM’s documentation describes its zero-data-retention mode as an OpenAI-compatible gateway that sends requests to partners operating inference capacity. It distinguishes this mode from its confidential inference mode using trusted execution environments. The documentation does not name individual partners.
PREM Confidential API’s privacy policy names PREM SA as its operator and lists Nebius B.V. as an inference subprocessor. It specifies transient processing of prompts and completions in the EU/UK, zero retention under a DPA and no training use.
PREM Confidential API privacy policyListed 20 Oct 2026 · PREM · Web pagePREM Confidential API’s privacy policy names PREM SA as its operator and lists Nebius B.V. as an inference subprocessor. It specifies transient processing of prompts and completions in the EU/UK, zero retention under a DPA and no training use.
PREM Confidential API is an AI inference platform operated by PREM SA in Lugano, Switzerland. It gives developers and businesses access to language models through an API and dashboard. Its confidential inference mode encrypts requests on the client and processes them inside hardware-isolated trusted execution environments. A separate OpenAI-compatible zero-data-retention mode routes requests to inference partners under contractual retention restrictions. Developers can integrate model responses into applications while choosing between these different privacy guarantees.
Prem AI launches CyberScan for continuous code security audits
Prem AI launched CyberScan in beta for continuous security reviews, initially targeting Bitcoin infrastructure. It supports resumable repository scans, prioritised findings, and integration with GitHub, APIs and MCP. ArkLabs and Breez tested it against production code before launch.
