Proton has been building a privacy-focused alternative to Big Tech’s productivity stack for years, covering email, cloud storage, VPN, and password management. Last year, the Swiss company added an AI chatbot to that lineup. Now, that chatbot is getting a serious upgrade.
Lumo 2.0, which Proton announced Tuesday, brings image handling, faster response times, and improved memory to a product that previously lagged behind competitors in raw capability. The update is available immediately, and the core version remains free.
The question Proton is really trying to answer here is whether you have to give up privacy to get a useful AI assistant. So far, most people have accepted that trade-off by defaulting to ChatGPT, Gemini, or Claude. Proton is betting a growing number of users won’t want to.
The biggest additions in version 2.0 are image-related. Users can now upload photos directly into Lumo for analysis or editing, and the chatbot can also generate images from text prompts. That puts it roughly in line with what tools like ChatGPT and Gemini already offer, which matters because feature parity is the baseline for anyone considering a switch.
Speed also gets a significant boost. Proton says Lumo 2.0 responds to most queries up to 76% faster than the previous version. There is also a new “thinking mode” for more complex or nuanced questions, similar to reasoning modes found in other modern AI tools.
On the organizational side, Proton has improved Projects, the workspace feature that lets users pull in documents and connect Lumo to other Proton products like Proton Mail and Proton Drive. Projects now support persistent memory that users control, meaning Lumo can remember preferences and context across separate conversations. That kind of continuity has become a standard expectation in AI tools, and its absence was a noticeable gap in the first version.
Proton CEO Andy Yen framed the update as proof that privacy and capability don’t have to be opposites. “Lumo 2.0 has been re-engineered from the ground up and the introduction of thinking mode gives it powerful new capabilities,” he said. “Lumo 2.0 demonstrates that users no longer need to choose between powerful AI capabilities and meaningful privacy protections.”
The privacy architecture behind Lumo is what actually sets it apart. Proton uses zero-access encryption, meaning data is encrypted both in transit and at rest, and only the user can access it. The company says no server-side logs of conversations are kept, so even Proton employees cannot read what users type. Proton also commits to never using customer data for AI training and never sharing it with third parties. That last point is a meaningful distinction from most major AI providers, which typically use conversations to improve their models unless users opt out.
This update arrives at a moment when concern about AI data practices is growing. Regulators in Europe have been scrutinizing how AI companies handle personal data, and users are becoming more aware of what happens to their conversations. Proton, based in Geneva and already well-known among privacy-conscious users, is positioned to capture some of that audience if Lumo can hold its own against more established tools.
In terms of day-to-day usefulness, Lumo 2.0 looks comparable to mainstream chatbots. It answers questions with similar depth and format to ChatGPT or Gemini. The gap that remained in the first version was mostly about features, not quality, and this update closes a good chunk of it.
Lumo is available in three tiers:
- Free – Public access with standard usage limits
- Plus – Paid tier with more resources and higher usage caps
- Professional – Paid tier with the most access, aimed at business users
For anyone already using Proton’s other products, adding Lumo to the mix now makes more practical sense than it did a year ago.