Proton Gives Lumo a Major AI Upgrade While Keeping Privacy at the Center

Proton has rolled out a major upgrade to Lumo, its privacy-oriented AI chatbot, adding image understanding, image generation, faster responses and a new reasoning feature designed for harder questions. The update marks a significant step for the Swiss company as it tries to prove that advanced AI tools do not have to rely on invasive data collection.

The refreshed version, Lumo 2.0, broadens the chatbot’s capabilities while keeping the company’s long-standing pitch intact: users should be able to use modern AI without giving up control over their information. Proton says the new release is available now to all users, with additional capacity and features reserved for paid plans.

What changed in Lumo 2.0

Proton’s latest release transforms Lumo from a text-first assistant into a more versatile AI product. The chatbot can now interpret uploaded images and generate new ones from prompts, placing it closer to the feature set offered by the biggest mainstream AI assistants.

That matters because image support has become a basic expectation in consumer AI. Users increasingly want one tool that can answer questions, summarize documents, inspect screenshots, draft visuals and help with everyday work tasks. Lumo 2.0 now fits that mold more closely than before.

Image recognition and generation arrive

The most visible addition is image handling. Users can upload photos or screenshots and ask Lumo to analyze them, describe them or help edit them. On the output side, the chatbot can also create images based on written prompts.

That combination puts Lumo in direct competition with the image features built into other major AI platforms. The practical use cases are straightforward: identifying objects in a photo, reviewing a chart, making a simple visual mockup or generating an illustration for a presentation draft.

For a company built on privacy messaging, the image update is also strategic. It signals that Proton is not positioning Lumo as a minimalist niche tool, but as a full-featured assistant capable of keeping up with the broader market.

Projects get persistent memory

Proton is also expanding the chatbot’s Projects feature, a workspace designed for users who want to combine AI with files from Proton’s ecosystem, including email and cloud storage. The biggest change here is user-controlled persistent memory.

In simple terms, that means Lumo can remember preferences over time across multiple conversations. The company says the memory feature is governed by the user, which is meant to preserve the privacy-first model while still making the assistant more useful.

This kind of memory has become increasingly common in chatbot products. It helps the AI avoid making users repeat instructions, remember formatting preferences and maintain continuity in longer workflows. But it also raises obvious concerns about how much a system stores, where it stores it and who can access it. Proton is making those issues part of its value proposition rather than trying to hide them.

Faster answers and a new thinking mode

Alongside the new features, Proton says Lumo 2.0 is significantly quicker than the previous version. The company says the chatbot responds to most questions up to 76% faster than before.

That speed boost is important because response time can shape how useful a chatbot feels in everyday use. Users are less likely to keep relying on a slow assistant, even if it is accurate. A faster model improves the experience for quick lookups, drafting tasks and iterative work.

A deeper mode for harder questions

The update also adds what Proton calls a “thinking mode,” intended for more complex requests. While the company did not provide detailed technical information, the feature appears to be aimed at more deliberate reasoning tasks, where the chatbot can spend more time working through a question before responding.

Reasoning-style modes have become a common feature in high-end AI assistants. They are typically used for analysis, multi-step problem solving, coding help or tasks that benefit from more structured output. Proton’s addition suggests Lumo is maturing beyond a privacy alternative into a product that aims to match user expectations shaped by OpenAI, Google and others.

Andy Yen, Proton’s founder and chief executive, said the chatbot was rebuilt substantially for this release and argued that the new version shows people do not have to trade strong AI features for privacy protections.

How Proton is trying to differentiate Lumo

Functionally, Lumo now looks more like the mainstream chatbots that dominate consumer AI. It answers questions in a conversational format, offers contextual responses and now supports images and memory. But Proton’s competitive angle is not simply feature parity. The company is betting that privacy can be a deciding factor in AI adoption.

That is a notable strategy in a market where many users are uneasy about what chatbots do with their prompts, files and personal details. Consumer AI tools often depend on collecting broad usage data to improve services, train models or personalize output. Proton is trying to move in the opposite direction.

Zero-access encryption as the core pitch

Proton says Lumo is built on what it describes as zero-access encryption architecture. According to the company, data is encrypted both when it is being transmitted and when it is stored, and only the user can access it.

The company further says it does not keep server-side logs of conversations, meaning Proton staff should not be able to read user chats. It also says customer data will not be used to train AI systems or passed to outside parties.

These assurances are central to Proton’s brand. The company made its name with encrypted email and privacy tools, and Lumo extends that identity into the AI market. In effect, Proton is arguing that the next generation of chatbots should inherit the same privacy expectations people already demand from secure messaging and cloud storage.

Why privacy matters more in AI than ever

AI chatbots are especially sensitive products because users often feed them personal documents, confidential work materials, legal questions, medical concerns and creative drafts. A chatbot that retains or repurposes that information can become a privacy liability very quickly.

That risk has pushed some users, businesses and regulators to look more closely at how AI vendors store, process and learn from customer data. Proton is leaning into that concern rather than treating it as a side issue. By making privacy part of the product itself, the company hopes to win users who would otherwise hesitate to use a general-purpose chatbot.

The challenge, of course, is that privacy alone does not guarantee adoption. Users still expect quality, speed and reliability. Lumo’s latest version appears designed to close that gap.

How Lumo compares with the big names

In day-to-day use, Proton’s chatbot now resembles mainstream products like ChatGPT and Gemini more than a stripped-down secure assistant. The responses appear broadly similar in length, structure and level of detail, at least based on Proton’s description of the product.

That creates a useful benchmark. If Lumo can deliver familiar chatbot performance while maintaining stronger data protections, it could appeal to privacy-conscious individuals, journalists, lawyers, researchers, startups and enterprises that are cautious about sensitive information entering third-party AI systems.

Where Lumo still stands apart

Despite the new features, Proton’s advantage is not based on scale. It does not have the same model ecosystem, distribution footprint or developer network as the biggest AI companies. Instead, it is trying to win on trust, product coherence and privacy guarantees.

That narrower pitch may be a strength. In a crowded market, AI products often blur together. Proton is giving Lumo a clearer identity: a capable assistant that is meant to be used without sacrificing user control.

Still, users will likely judge Lumo on a combination of factors:

• how accurately it handles text and image prompts
• whether its memory feature is genuinely useful
• how quickly it responds under load
• how transparent Proton is about data handling
• whether its paid plans offer enough value to justify the cost

Availability and pricing

Lumo 2.0 is available now. Proton is keeping a free public version in place, while also offering Plus and Professional tiers that provide higher usage limits and greater access to resources.

That pricing structure is standard across the AI sector. Free tiers help attract users and build awareness, while paid plans monetize frequent usage and business demand. Proton appears to be following the same model, but with an added privacy hook that may make the premium tiers more attractive to security-conscious customers.

For many users, the key question will be whether the free version is enough for everyday needs. For others, especially professionals handling sensitive documents, the appeal of a paid privacy-first AI assistant may be easier to justify.

Why this update matters for the wider AI market

Lumo’s upgrade arrives at a moment when AI assistants are becoming more capable and more embedded in daily workflows. The competitive bar has risen. A chatbot no longer needs to merely answer questions; it must handle images, retain context, operate quickly and support more complex reasoning.

At the same time, public anxiety about data use continues to shape the market. The more powerful AI systems become, the more users want to know what happens to their prompts, uploads and account history. Proton is exploiting that tension.

Lumo 2.0 suggests a possible direction for the next phase of consumer AI: feature-rich assistants that are designed from the start around strict data controls. If that model gains traction, it could pressure other providers to offer clearer privacy settings or more limited data retention policies.

A test case for privacy-first AI

Proton has a strong reputation among privacy-minded users, but Lumo is an important test of whether that reputation can carry into a fast-moving AI category. The product now offers enough mainstream capabilities to be taken seriously on its own merits.

If the chatbot proves accurate, fast and genuinely private, it may serve as a template for a different kind of AI business: one where trust is not an afterthought but the core reason to choose the product.

Timeline of Lumo’s evolution

What users should know before trying it

Anyone considering Lumo 2.0 will want to look past the marketing language and focus on practical trade-offs. The upgrades make the assistant more capable, but the privacy claims remain the defining feature.

Users should pay attention to what kind of information they are comfortable sharing, whether memory is enabled by default in their workflows, and how the product behaves across free and paid tiers. As with any AI tool, the best approach is to assume that convenience and control must be balanced carefully.

Proton is betting that more people are ready for that balance than ever before. With Lumo 2.0, it is making a clearer case that high-end AI and strong privacy protections do not have to exist in separate markets.

The bottom line

Lumo 2.0 is more than a routine update. It gives Proton’s chatbot the kind of capabilities users now expect from a serious AI assistant, while reinforcing the company’s insistence that private data should stay private.

Whether that combination proves compelling enough to stand out in a crowded chatbot market will depend on execution. But the message is now unmistakable: Proton wants Lumo to compete not by copying the surveillance-heavy playbook of the AI industry, but by offering a different one entirely.

superintelligencenews@gmail.com