Proton Ships Lumo 2.0 With Image Generation and Private Search

*Proton rebuilt its AI assistant on a new architecture that adds memory, image tools, and faster responses while keeping zero-access encryption.*

Proton released Lumo 2.0 on June 30. The update introduces image generation, image recognition, memory, and private web search to the privacy-first chatbot first shipped last summer.

The company rebuilt the service on a fresh architecture. It now offers two response modes: Fast for quick answers and Thinking for multi-step reasoning. Proton states that everyday queries run up to 76 percent faster than in version 1.4.

New capabilities

Users can generate images and upload images for analysis. Memory lets the assistant retain context across sessions. Web search operates under the same no-logs, no-sharing rules that define the rest of Proton’s products.

All new functions run with zero-access encryption, so Proton cannot read user data or prompts. The company says this is the largest single upgrade since launch.

Privacy constraints remain

Lumo 2.0 keeps the original design choices. No conversation logs are stored, data is not shared with third parties, and encryption prevents server-side inspection. Proton positions these limits as the main distinction from mainstream assistants.

The sources report no changes to pricing or availability. The update applies to existing users without requiring new accounts.

Reactions

No third-party reviews or competitor statements appear in the source material. Proton’s own announcements emphasize capability gains inside the same privacy envelope.

Why it matters

Engineers and founders who already route mail, calendar, and storage through Proton now have an on-platform AI option that does not send prompts to outside model providers. The 76 percent speed claim and the addition of image tools narrow the practical gap with general-purpose chatbots while preserving the encryption boundary. Whether the memory feature stores data only on-device or in encrypted cloud storage is not detailed in the published material, so users must test the behavior themselves.

The move shows Proton treating AI as another service that must obey its core encryption model rather than an exception. For customers already inside the Proton ecosystem, the update removes one reason to leave for another assistant.

---

Sources:

• 9to5Mac
• MacRumors
• Neowin

{
  "excerpt": "Proton ships Lumo 2.0 with image generation, memory, private search, and a new architecture that runs everyday queries up to 76 percent faster.",
  "suggestedSection": "ai",
  "suggestedTags": ["proton", "ai-assistant", "privacy"],
  "imagePrompt": "An abstract arrangement of layered translucent panels and soft geometric forms suggesting layered data and encrypted memory, set against a quiet studio backdrop with subtle light gradients. muted color palette, cinematic lighting, 16:9"
}