Global privacy and security company ExpressVPN today announced ExpressAI, a new private-by-design AI platform designed for people who want the power of modern AI without being profiled or having their prompts and files fed back into training pipelines.

As more people use AI to draft emails, summarise documents, and troubleshoot everyday problems, they’re also increasingly sharing more sensitive content, like work files, financial questions, and personal details. Too often, that means choosing between powerful AI features and personal privacy. ExpressAI is built to remove that trade-off, ensuring user data remains encrypted and unreadable to anyone, including ExpressVPN.

Unlike most existing AI platforms that still process data on traditional servers, ExpressAI brings groundbreaking secure computing enclave technology to everyday users. This approach ensures that conversations are decrypted only within a secure enclave, a cryptographically isolated environment that is inaccessible to the host system and infrastructure operators.

“With ExpressAI, ExpressVPN is effectively extending its long-held stance on traffic protection to AI interactions: The best way to protect user data is not to collect it in the first place,” said Shay Peretz, COO at ExpressVPN. “We’re not just making privacy claims – we’re proving it with cryptographic guarantees. With our enclave architecture, your messages exist in a secure, isolated environment that even we can’t access. This is the future of reliable, private AI.”

A Privacy-first AI Architecture

ExpressAI brings together privacy, model choice, and real usability in a single product. ExpressAI delivers a stack of privacy-focused features that set the platform apart from traditional AI chat tools:

• Zero-access: ExpressAI applies zero-knowledge, end-to-end encryption that ensures only users can decrypt their data. Neither ExpressVPN nor any third party can view or access any data
• Zero training: User conversations, prompts, and files are not fed back into model training
• Private file handling: Files are protected using the same user-controlled encryption approach as chat history, helping ensure stored content remains private and inaccessible to anyone else.
• Ghost mode: Conversations can be configured to automatically delete.
• Encrypted vault: Users set their own password to create an encrypted vault that only they can unlock, ensuring chat history remains private and accessible only to them.

At launch, ExpressAI offers a suite of five general-purpose models to cover everyday tasks:

• GPT OSS 120B: A powerful open-weight model by OpenAI for everyday reasoning and writing. Excels at drafting, summarisation, Q&A, and general-purpose tasks.
• DeepSeek R1 Distill 32B: Your go-to model for problems that require careful thinking. Excels at multi-step reasoning, logic puzzles, research analysis, and anything that benefits from thinking before answering.
• Qwen2.5-VL 32B: See and understand the world around you. Excels at reading documents, extracting data from images, answering questions about charts, and analysing diagrams.
• Qwen3.5 35B-A3B: A smart, efficient model for long and complex tasks. Excels at coding, multi-step reasoning, and multilingual tasks.
• Nemotron 12B: A compact powerhouse by NVIDIA for technical work. Excels at solving math problems, generating code, and tackling complex reasoning tasks.

ExpressAI features a convenient side-by-side comparison view, allowing users to run the same prompt across multiple models simultaneously and benefit from a range of perspectives and reasoning styles.

It also offers transparent credit tracking for clear visibility into usage. For chat history, ExpressAI uses user-controlled encryption: users set a primary password, and only that password can decrypt past conversations, ensuring chat history remains unreadable to anyone else, including ExpressVPN.

Independently audited by Cure53

ExpressAI has been independently audited by Cure53, a leading cybersecurity firm known for its rigorous security assessments.

The audit, conducted in February and March 2026, included penetration testing and a source code review across ExpressAI’s frontend and backend systems, as well as its cryptography, key management, and underlying infrastructure.

“Cure53 concludes that the product meets its stated privacy objectives by providing modern AI capabilities within confidential computing enclaves, where user interactions are processed in cryptographically isolated contexts,” said Cure53.

All identified security vulnerabilities by Cure53 were addressed and remediated ahead of launch. The full report is available here.

“People are already turning to AI for high-stakes, personal conversations – from health questions to financial decisions. Whether you’re accessing a bank account online or discussing private matters with a professional, you expect strong privacy protections. But those protections don’t automatically carry over to everyday AI chats,” said Peretz. “That gap has left many users uneasy about how their data is handled. ExpressAI was built to remove that fear and show that private, trustworthy AI can be part of the tools people use every day.”

Privacy-by-design has always been core to ExpressVPN. With ExpressAI, the company is applying that same privacy-first architecture beyond VPN to the everyday tools people use most, including passwords, messaging, and now AI.

ExpressAI reflects ExpressVPN’s evolution from a single, core product into a broader privacy ecosystem. The company’s vision is to bring its uncompromising privacy and security principles to more digital tools people use every day.

ExpressAI will roll out in phases, with new and existing ExpressVPN Pro plan users receiving first access. For more information, visit https://www.expressvpn.com/expressai.