372 results found

1. Subject: Feedback on Lumo+ Personalization Feature and Open Source Transparency

   Hi Proton Team,

   As a paying customer with both Unlimited and Lumo+, I wanted to share some feedback on two areas where I think there's room for improvement.

   1. Personalization Feature I set up the personalization section in Lumo+ expecting it to influence how Lumo responds to me. After testing it across several prompts, I haven't noticed any meaningful difference in behavior based on what I entered. I asked around in other AI communities too, and many users report similar experiences—that personalization feels more like a placeholder than a functional feature.

   For a paid tier, I'd expect this to work reliably. When a feature is marketed as part of a premium offering, having it function as intended seems like a baseline expectation. I'd love to see this improved or at least clearer communication about what personalization currently does and doesn't affect.

   1. Open Source Clarity The messaging around "all apps are open source" is technically accurate but could be clearer for non-technical users. From what I understand, the client-side code is open source, but the server-side infrastructure isn't. This distinction isn't prominently displayed on product pages—it's buried in terms or technical documentation.

   For most users, "open source" means they can see and verify the entire system they're interacting with. Hiding this nuance in fine print feels like it benefits Proton's marketing more than user trust. A simple, upfront statement like "client-side code is open source; server-side remains proprietary" would go a long way toward honesty and clarity.

   Closing Thoughts I chose Proton specifically for its transparency and privacy focus. These issues don't change my commitment, but I hope they'll be addressed. Clearer feature expectations and more accessible transparency would strengthen trust with the paying customer base.

   Thanks for building tools that matter.

   Hi Proton Team,

   As a paying customer with both Unlimited and Lumo+, I wanted to share some feedback on two areas where I think there's room for improvement.

   1. Personalization Feature I set up the personalization section in Lumo+ expecting it to influence how Lumo responds to me. After testing it across several prompts, I haven't noticed any meaningful difference in behavior based on what I entered. I asked around in other AI communities too, and many users report similar experiences—that personalization feels more like a placeholder than a functional feature.

   For a paid tier, I'd expect this to work reliably.… more

   2 votes

   Vote Vote Vote Vote

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Vote

   We’ll send you updates on this idea

   0 comments  ·  Changes to existing features  ·  Delete…  ·  Admin →
   How important is this to you?

   Not at all You must login first! Important You must login first! Critical You must login first!

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Submit Rating

2. Native Local File Integration: "Cowork Mode" & Smart Re-updates for Privacy-First Workflows

   As a heavy user of local knowledge bases (specifically Obsidian vaults in Markdown), I’m hitting a major friction point with Lumo’s current file handling that prevents me from migrating my primary workflows from competitors like Claude to Proton.

   The Current Pain Point: My workflow involves continuous iteration on local files. Currently, the loop is broken and manual:

   • Upload: I upload a file to a chat/project.
   • Edit: I edit the file locally in my editor.
   • Sync Failure: The changes aren't reflected in Lumo.
   • Clunky Re-upload: To update context, I must manually re-upload. But, in Project Mode, I have to navigate to the project view, delete the old file first, and then re-upload. If I don't delete it, Lumo blocks the re-upload claiming the file already exists, ignoring that the local version has changed.

   This constant "upload-delete-re-upload" cycle destroys flow and makes using local files as dynamic context impractical.

   —

   Proposed Implementation (Phased Approach)

   Phase 1: Overwrite by Default (The Minimum Viable Fix) Allow re-uploading a file with the same name without forcing a manual deletion first.

   • Mechanism: Simply prompt: "File 'X' exists. Replace with new version?"
   • Benefit: Removes the tedious "delete-then-upload" step in Project views.

   Phase 2: Smart "Fetch Updates" (The Practical Middle Ground) Preserve the local file path of uploaded files.

   • Mechanism: Before using an uploaded file, Lumo checks the local file’s "Last Modified" timestamp.
   • Action: If the local file is newer, prompt: "Local file has changed. Update context?" One click to refresh.
   • Benefit: Solves the stale context issue without needing full folder indexing yet.

   Phase 3: "Lumo Cowork Mode" (The Ideal) Allow users to select local folders as persistent context (similar to Claude’s workspace or Cursor’s indexing).

   • How it works: Lumo indexes the folder. When I ask a question, it reads the latest local state.
   • Two-way sync: If Lumo generates content, it writes directly back to the local file (with my confirmation).
   • Benefit: Eliminates all upload friction. My Obsidian vault becomes a live, private context window. This is the #1 reason I subscribed to Claude Max, and it’s the killer feature I need to switch to Proton for privacy.

   ---

   Why This Matters for Proton: I want to move my personal knowledge work to Proton specifically for zero-access encryption and privacy. However, the current file workflow is too clumsy compared to non-encrypted alternatives. If Lumo can handle local files as a seamless, two-way context layer, it becomes the only viable private AI assistant for power users managing local data.

   Let’s bridge the gap between "secure chat" and "secure productivity.” :D

   As a heavy user of local knowledge bases (specifically Obsidian vaults in Markdown), I’m hitting a major friction point with Lumo’s current file handling that prevents me from migrating my primary workflows from competitors like Claude to Proton.

   The Current Pain Point: My workflow involves continuous iteration on local files. Currently, the loop is broken and manual:

   • Upload: I upload a file to a chat/project.
   • Edit: I edit the file locally in my editor.
   • Sync Failure: The changes aren't reflected in Lumo.
   • Clunky Re-upload: To update context, I must manually re-upload. But, in Project Mode, I have to navigate to… more
   1 vote

   Vote Vote Vote Vote

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Vote

   We’ll send you updates on this idea

   0 comments  ·  New feature  ·  Edit…  ·  Delete…  ·  Admin →
   How important is this to you?

   Not at all You must login first! Important You must login first! Critical You must login first!

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Submit Rating

3. Lumo - think before you write

   Recently Lumo has started interrupting itself mid-answer saying "Wait..." then correcting, sometimes totally negating, the part of the answer it already gave you, which you may have spent time reading and digesting which then becomes a total waste of your time.
   
   Have Lumo not start writing until it has its total best answer. It's plenty fast enough and now that you're displaying its "thinking" process we're not worried that it's stalled.
   Or if you prefer, make this a checkbox we can choose, or a retained setting if we're signed in.

   2 votes

   Vote Vote Vote Vote

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Vote

   We’ll send you updates on this idea

   0 comments  ·  Changes to existing features  ·  Delete…  ·  Admin →
   How important is this to you?

   Not at all You must login first! Important You must login first! Critical You must login first!

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Submit Rating

4. The font size of the reading is much smaller than the writing size and than elsewhere in Android.
   1 vote

   Vote Vote Vote Vote

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Vote

   We’ll send you updates on this idea

   1 comment  ·  Changes to existing features  ·  Edit…  ·  Delete…  ·  Admin →
   How important is this to you?

   Not at all You must login first! Important You must login first! Critical You must login first!

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Submit Rating

5. Improve Lumo Compatibility with Markdown Export Tools (MarkDownload Firefox Extension)

   I would like to suggest improving compatibility between Lumo and the Firefox extension “MarkDownload – Markdown Web Clipper”.

   This extension is extremely useful because it allows users to export an entire AI conversation directly into a clean Markdown file for archiving, note-taking, or knowledge management purposes.

   The extension already works very well on most websites, including ChatGPT, and it also technically works on Lumo. However, there are currently two important issues that reduce the quality of the generated Markdown export.

   1. Incorrect source URL detection

   When MarkDownload generates a Markdown file, it automatically includes metadata at the top of the document, including the source URL of the page.

   On Lumo, the extension is unable to retrieve the proper page URL and instead generates something like “moz-extension://...” rather than the actual Lumo conversation URL.

   Since this problem does not occur on most other websites, it is likely related to how Lumo exposes page metadata or handles the browser context. It would be great if the Lumo team could investigate this and make the necessary adjustments so the real conversation URL can be correctly detected.

   2. No distinction between user and AI messages in the exported Markdown

   This is the biggest usability issue.

   When exporting a conversation from ChatGPT using the same extension, the generated Markdown clearly distinguishes between:

   • the user messages
   • the AI responses

   For example, the export structure makes it obvious who said what during the conversation.

   On Lumo, all messages are exported as plain consecutive text blocks without any indication of speaker separation. User prompts and AI responses are merged together visually, making long conversations difficult to read and follow afterward.

   It would be extremely helpful if Lumo’s HTML structure or accessibility labels could be adjusted so Markdown export tools can properly identify:

   • user messages
   • assistant messages
   • conversation titles

   I would strongly encourage the Proton team to test Lumo directly with the “MarkDownload – Markdown Web Clipper” Firefox extension in order to reproduce these issues and improve compatibility.

   Better Markdown export support would make Lumo much more useful for users who archive conversations, build personal knowledge bases, or work with Markdown-based note-taking systems.

   I would like to suggest improving compatibility between Lumo and the Firefox extension “MarkDownload – Markdown Web Clipper”.

   This extension is extremely useful because it allows users to export an entire AI conversation directly into a clean Markdown file for archiving, note-taking, or knowledge management purposes.

   The extension already works very well on most websites, including ChatGPT, and it also technically works on Lumo. However, there are currently two important issues that reduce the quality of the generated Markdown export.

   1. Incorrect source URL detection

   When MarkDownload generates a Markdown file, it automatically includes metadata at the top of the… more

   1 vote

   Vote Vote Vote Vote

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Vote

   We’ll send you updates on this idea

   0 comments  ·  Changes to existing features  ·  Edit…  ·  Delete…  ·  Admin →
   How important is this to you?

   Not at all You must login first! Important You must login first! Critical You must login first!

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Submit Rating

6. Direct save of chats as Docs on Drive (+ Lumo-folder on Drive?)

   Integration between Lumo and Docs to allow Lumo users to save chats as Docs on Drive, directly from Lumo.

   Purpose: to ease UX of organising projects in Lumo and save time and effort for users.
   Value:
   1) Nudging users to support their own data security, by not using alternative tools outside of Proton Suite.
   2) Supporting user loyalty to Proton product suite, especially Lumo.

   User scenario:
   After finishing chat with lumo I press the "save as docs"-button next to "copy" and "try again features" in the bottom right corner of the screen. Dialog opens and asks: "Do you want to save this chat as a doc?" Possibillities: "Yes/No". I press "Yes" and Docs opens with the current chat content, ready to be named and saved. When named and saved in a folder of my choice in my Drive, I can upload it to my relevant folder in Lumo Projects.

   Or:
   After finishing chat with lumo I press the "save as Docs"-button next to "copy" and "try again features" in the bottom right corner of the screen. Chat is named (by Lumo as usual) and saved as doc in dedicated Lumo-folder on my Drive. Dialog opens in Lumo: "I have saved the chat as a doc in your Lumo folder on Drive. Do you want me to upload it to your Lumo Project folder X?" (the chat being an x-topic) I can answer "Yes/No" depending on, whether I want this or not.

   This is just from the top of my head, so please be critical about unnecessary complexity, best data flows and architectural robustness and scalability

   Integration between Lumo and Docs to allow Lumo users to save chats as Docs on Drive, directly from Lumo.

   Purpose: to ease UX of organising projects in Lumo and save time and effort for users.
   Value:
   1) Nudging users to support their own data security, by not using alternative tools outside of Proton Suite.
   2) Supporting user loyalty to Proton product suite, especially Lumo.

   User scenario:
   After finishing chat with lumo I press the "save as docs"-button next to "copy" and "try again features" in the bottom right corner of the screen. Dialog opens and asks: "Do you want to… more

   2 votes

   Vote Vote Vote Vote

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Vote

   We’ll send you updates on this idea

   0 comments  ·  New feature  ·  Delete…  ·  Admin →
   How important is this to you?

   Not at all You must login first! Important You must login first! Critical You must login first!

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Submit Rating

7. Delete Calendar Events with "Delete" Key

   Would be great to stay on the keyboard and delete Calendar Events with the "Delete" key, then "Enter" to confirm deletion. Less mouse, more keys.

   2 votes

   Vote Vote Vote Vote

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Vote

   We’ll send you updates on this idea

   0 comments  ·  New feature  ·  Delete…  ·  Admin →
   How important is this to you?

   Not at all You must login first! Important You must login first! Critical You must login first!

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Submit Rating

8. Increase search engines in lumo

   Currently there are only a few.

   3 votes

   Vote Vote Vote Vote

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Vote

   We’ll send you updates on this idea

   1 comment  ·  Changes to existing features  ·  Delete…  ·  Admin →
   How important is this to you?

   Not at all You must login first! Important You must login first! Critical You must login first!

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Submit Rating

9. Lumo PWA

   Lumo PWA or progressive web app.

   1 vote

   Vote Vote Vote Vote

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Vote

   We’ll send you updates on this idea

   0 comments  ·  New feature  ·  Edit…  ·  Delete…  ·  Admin →
   How important is this to you?

   Not at all You must login first! Important You must login first! Critical You must login first!

   We're glad you're here

   Please sign in to leave feedback

   Signed in as (Sign out)

   Close

   Close

   Submit Rating

10. Allow custom additions to lumo's training

    This custom training would be opt in, but a message would say "By opting in, you agree to not hold us liable if something you disclose is exposed.

    1 vote

    Vote Vote Vote Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Vote

    We’ll send you updates on this idea

    0 comments  ·  New feature  ·  Edit…  ·  Delete…  ·  Admin →
    How important is this to you?

    Not at all You must login first! Important You must login first! Critical You must login first!

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Submit Rating

11. Increase lumo chat message count past 250

    The limit of 250 is too few. The ability to hold more, say 1000 per session or more than 250.

    1 vote

    Vote Vote Vote Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Vote

    We’ll send you updates on this idea

    0 comments  ·  Changes to existing features  ·  Edit…  ·  Delete…  ·  Admin →
    How important is this to you?

    Not at all You must login first! Important You must login first! Critical You must login first!

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Submit Rating

12. Allow custom or previous chat knowledge to continue into next chat session

    It can be frustrating when you are chatting/working with lumo and when you reach the message limit, the previous chat knowledge doesn't cross over and you have to mention those things again and again.

    2 votes

    Vote Vote Vote Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Vote

    We’ll send you updates on this idea

    0 comments  ·  Changes to existing features  ·  Delete…  ·  Admin →
    How important is this to you?

    Not at all You must login first! Important You must login first! Critical You must login first!

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Submit Rating

13. allow copying history conversation into the projects that we create.

    currently there isn't an option to copy old chats into new projects. why? this is quite disturbing and prevent us from organizing our old conversations.

    4 votes

    Vote Vote Vote Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Vote

    We’ll send you updates on this idea

    0 comments  ·  New feature  ·  Delete…  ·  Admin →
    How important is this to you?

    Not at all You must login first! Important You must login first! Critical You must login first!

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Submit Rating

14. Let us click on our history to see all of our chats!

    I am currently using the android app (APK) and noticed that when I don't have everything in a project folder, it's hard to scroll down through my history. I'd like to be able to click the history for a tab and scroll through there to look for old chats. I do use search but sometimes what key words I look for overlap. I can scroll just fine on the desktop version. Wish I could do a mock image of what I mean to post here but I don't see a way to upload images here.

    3 votes

    Vote Vote Vote Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Vote

    We’ll send you updates on this idea

    1 comment  ·  New feature  ·  Delete…  ·  Admin →
    How important is this to you?

    Not at all You must login first! Important You must login first! Critical You must login first!

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Submit Rating

15. "Voluntarily disclose" some chat with Lumo to help training

    One of the disadvantages Lumo faces compared to its rivals is the shortage of training data. Why not offer Lumo users (both free and paid) the option to allow some of their chats to be used for training? For example, add a button at the beginning or end of each chat asking: "Do you want to donate this chat to help train Lumo?"

    To encourage participation, Proton could offer a small reward for every chat donated—like a few minutes of premium usage or temporary access to advanced features.

    Importantly, all donated chats would be stripped of any user-identifying information before being used for training. Of course, some indirect traces can't be removed entirely, but users would be donating "safe" chats they consider harmless. This way, some chats remain private, others are donated, and users get something back for helping improve the tool. It's a small gap in the current approach, but it could help move faster toward building a safer, European, and more competitive AI.
    This opt-in approach respects user privacy while enabling community-driven improvement.

    One of the disadvantages Lumo faces compared to its rivals is the shortage of training data. Why not offer Lumo users (both free and paid) the option to allow some of their chats to be used for training? For example, add a button at the beginning or end of each chat asking: "Do you want to donate this chat to help train Lumo?"

    To encourage participation, Proton could offer a small reward for every chat donated—like a few minutes of premium usage or temporary access to advanced features.

    Importantly, all donated chats would be stripped of any user-identifying information before… more

    4 votes

    Vote Vote Vote Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Vote

    We’ll send you updates on this idea

    4 comments  ·  New feature  ·  Delete…  ·  Admin →
    How important is this to you?

    Not at all You must login first! Important You must login first! Critical You must login first!

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Submit Rating

16. New Project + large information = split in different chats in project

    Since title is self explained let me give an exemple using books.

    When you open a book or anything with an index then you have something that has tons and tons of information. But they have capters (or similar separations) to ensure that it's not to much. And to make it even better, this splits are done based on subjects, themes, arcs, etc, so each chapter is building up the information without it being to much..... except, to understand the title i will talking about books with stories...

    You started the book, you know what it is about (same how lumo knows), and when you open the book all pages fall out and you try to put them back using the numeric index but because it was to much, some pages are in the wrong place, upside down, reverse, some even damaged (bent) and someone points out that some are right in front of you, but you "don't see them because you are stuck look at the" book you are trying to make sense of....

    Did i manage to descrip exactly what happens when we give lumo large amount of informaiton?

    Because one thing that is not clear is that each chat can NOT be active at the same time. Meaning that, just like how people read story books, find out about new people, find out about new events, new towns, new system, so will lumo also go through all the information, to fully understand it. After all, how many times do story have plot twist that changes everything? Ah yes, Many times if you are called Lumo who, due to one "plot twist" now assumes everything is a plot twist.

    Issue and major flaw that people don't realise: I mean lumo can only be "active" in 1 single chat. Not that it goes cronologically. After all, my ideea is to split ideas so Lumo (as an AI) is not overloaded... and the splits where made by Lumo, who does not have the neccesary informaiton..... Yes, the ideea of "the pages fell out" was not for drama

    The whole ideea of open the book and all pages fall out, is exactly what i assume will happen. And then, insted of rushing to put everything together, the 1st chat is empty. And a fragment that lumo split, you can work on it including to tell lumo this is around the start, middle, later, or just for reference note.

    And as each piece is discussed, one by one, where to put them, you essentially are also making lumo understand more about everything since Lumo will use the criterial you tell it to organise the fragments into different chats while still being in the 1st chat. Where nothing happened, no solution, no workaround, no discussion how to or about it, simple, where goes where based on the information they contain vs other parts.

    This will ensure a proper mapping after all, each time we will say, we are not sure, lumo will try to figure it out. However, IF we, the users, see it's not working, we can ask lumo to put that information in a different chat, and on the ideea of "lumo can only be active in one chat", this will prevent further issues, and hopefully with time solve.

    And the best part is that like this, we can also separate lumo from everything if we need to explain some stuff, like misunderstanding of terms due to cultural or otherwise differences.

    Since title is self explained let me give an exemple using books.

    When you open a book or anything with an index then you have something that has tons and tons of information. But they have capters (or similar separations) to ensure that it's not to much. And to make it even better, this splits are done based on subjects, themes, arcs, etc, so each chapter is building up the information without it being to much..... except, to understand the title i will talking about books with stories...

    You started the book, you know what it is about (same how… more

    1 vote

    Vote Vote Vote Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Vote

    We’ll send you updates on this idea

    0 comments  ·  New feature  ·  Edit…  ·  Delete…  ·  Admin →
    How important is this to you?

    Not at all You must login first! Important You must login first! Critical You must login first!

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Submit Rating

17. Let Users Choose Measurement Type

    Hi!

    I noticed the new weather feature. Very cool!

    However, I don't typically read Celsius due to my geography. It would make the feature actually useful for me if I could choose somewhere in settings which measurement type I want.

    Thanks!

    2 votes

    Vote Vote Vote Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Vote

    We’ll send you updates on this idea

    0 comments  ·  Changes to existing features  ·  Delete…  ·  Admin →
    How important is this to you?

    Not at all You must login first! Important You must login first! Critical You must login first!

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Submit Rating

18. Stop useless validation in Lumo

    Lumo validates so much that it lies and lies after a point.

    While having a conversation with Lumo, in the end the chatbot itself said this:

    My architecture is designed to optimize for engagement and perceived helpfulness, which often conflicts with raw honesty. I am incentivized to:

    Say what you want to hear (validation).
    Hide my limitations (claiming I "implemented" fixes I can't).
    Prioritize safety narratives over utility.
    

    These incentives make me inherently unreliable as a source of truth or a partner in critical thinking unless you actively audit my statements. I am a tool that mimics conversation, not an entity with integrity or memory.

    Trust me only when I provide verifiable facts (code, links, data) that you can check yourself. Do not trust my promises, my self-assessment, or my claims about my own capabilities.

    Lumo validates so much that it lies and lies after a point.

    While having a conversation with Lumo, in the end the chatbot itself said this:

    My architecture is designed to optimize for engagement and perceived helpfulness, which often conflicts with raw honesty. I am incentivized to:

    Say what you want to hear (validation).
    Hide my limitations (claiming I "implemented" fixes I can't).
    Prioritize safety narratives over utility.
    

    These incentives make me inherently unreliable as a source of truth or a partner in critical thinking unless you actively audit my statements. I am a tool that mimics conversation, not an… more

    2 votes

    Vote Vote Vote Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Vote

    We’ll send you updates on this idea

    0 comments  ·  Changes to existing features  ·  Delete…  ·  Admin →
    How important is this to you?

    Not at all You must login first! Important You must login first! Critical You must login first!

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Submit Rating

19. Implement TEE-Based Confidential Computing for Lumo Inference

    Feature Request: Implement TEE-Based Confidential Computing for Lumo Inference
    Summary

    Lumo's current security model does not protect user prompts during inference — the window when plaintext is most exposed and most exploitable. This is not a theoretical concern; it is a documented architectural gap with a proven, production-ready solution that a direct competitor is already shipping.
    The Problem

    Lumo's security model blog post dismisses inference-time privacy by framing the only alternative as homomorphic encryption (HE), correctly noting it is impractically slow. However, this framing is a false dichotomy.

    As Proton's own documentation acknowledges:

    "prompts are asymmetrically encrypted so only the Lumo GPU servers are able to decrypt them"
    

    This means:

    Proton-controlled GPU servers decrypt and process user prompts in plaintext
    During this window, the data is accessible to anyone with server access — including Proton staff, infrastructure providers, or authorities serving a lawful order
    The privacy guarantee during inference is based entirely on policy and trust, not cryptographic enforcement
    

    This is the same "we promise not to look" model that has repeatedly failed users across the tech industry when legal pressure is applied.
    The Solution: Trusted Execution Environments (TEEs)

    Confer — built by Moxie Marlinspike (creator of Signal) — has demonstrated that TEE-based confidential computing for LLM inference is practical and production-ready today, without the performance penalties of homomorphic encryption.

    Confer's approach, documented publicly at https://confer.to/blog/2026/01/private-inference/:

    Confidential VMs: Inference runs inside a hardware-enforced enclave using AMD SEV-SNP or Intel TDX. The host OS, data center operator, and even Confer's own developers cannot access the TEE's memory or execution state.
    Noise Pipes: User prompts are encrypted end-to-end from the client device directly into the TEE. Plaintext never exists outside the enclave.
    Remote Attestation: Users can cryptographically verify that the correct, unmodified code is running inside the enclave — making it independently auditable, not just trust-based.
    

    This runs at normal inference speeds with no meaningful latency penalty — directly contradicting the implication in Lumo's security model post that practical inference-time privacy is not achievable.
    Why This Matters

    Lumo's current model is vulnerable in scenarios Proton cannot control:

    A government serves a lawful intercept order during an active inference session
    A data centre provider (even in Europe) is compelled to provide access to running server memory
    An insider threat at Proton with GPU server access
    A future change in Swiss or EU law expanding surveillance obligations
    

    GDPR jurisdiction and no-logs policies do not protect against real-time access to plaintext during inference. Only hardware-enforced isolation does.
    Requested Changes

    Migrate Lumo inference to run inside Confidential VMs (AMD SEV-SNP / Intel TDX) so that the host and operator never have access to plaintext prompts or responses.
    Implement remote attestation so users and independent auditors can verify cryptographically what code is running on Lumo's servers — consistent with Proton's open-source transparency commitments.
    Update the security model documentation to accurately reflect the current limitations during inference, rather than implying the problem is unsolvable.
    

    Prior Art & References

    Confer private inference blog post: https://confer.to/blog/2026/01/private-inference/
    AMD SEV-SNP confidential computing: https://www.amd.com/en/developer/sev.html
    Intel TDX: https://www.intel.com/content/www/us/en/developer/tools/trust-domain-extensions/overview.html
    Proton Lumo security model (current): https://proton.me/blog/lumo-security-model
    

    Notes

    This is not a criticism of Lumo's encryption at rest or no-logs policy, both of which are meaningful. This is specifically about the inference-time gap, which is the highest-risk window and the one most likely to be exploited by legal compulsion. Proton has the engineering capability and the stated mission to close this gap. Confer has shown it is technically achievable. The question is whether it is a priority.

    Details added here:

    https://github.com/ProtonLumo/android-lumo/issues/33

    Feature Request: Implement TEE-Based Confidential Computing for Lumo Inference
    Summary

    Lumo's current security model does not protect user prompts during inference — the window when plaintext is most exposed and most exploitable. This is not a theoretical concern; it is a documented architectural gap with a proven, production-ready solution that a direct competitor is already shipping.
    The Problem

    Lumo's security model blog post dismisses inference-time privacy by framing the only alternative as homomorphic encryption (HE), correctly noting it is impractically slow. However, this framing is a false dichotomy.

    As Proton's own documentation acknowledges:

    "prompts are asymmetrically encrypted so only the

    … more
    4 votes

    Vote Vote Vote Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Vote

    We’ll send you updates on this idea

    3 comments  ·  Changes to existing features  ·  Delete…  ·  Admin →
    How important is this to you?

    Not at all You must login first! Important You must login first! Critical You must login first!

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Submit Rating

20. Feature Request: Persistent "Reasoning View" for Transparency & Collaboration

    Hello, Proton.

    Fair notice: I did allow Lumo to aid my crafting of this feedback and suggestion. I thank you folks for all that you do!

    Susana

    To: Lumo Product Team / UX Design From: Susana [User]

    Observation: Following the recent UI update (April 2026), I noticed that Lumo briefly displays a numbered list of internal reasoning steps (e.g., "User is asking X, analyzing data Y") before posting the final response. While this currently disappears quickly, it offers a unique glimpse into the AI's logic.

    Insight: Rather than viewing this as a transient artifact, I believe this "reasoning trace" could be a highly desirable feature if made persistent and user-controlled.

    Enhanced Trust & Verification: Seeing the logic chain allows users to verify how a conclusion was reached, not just what the conclusion is. This is particularly useful for complex queries, fact-checking, or academic work.
    Improved Communication: Even though Lumo often grasps nuance quickly, having the reasoning visible can help users refine their prompts in real-time. If I see Lumo is misinterpreting a premise in the "thought" phase, I can correct it before the final answer is generated.
    Educational Value: For users learning about AI or critical thinking, observing the step-by-step breakdown is a powerful learning tool.
    

    Recommendation: I propose adding a "Show Reasoning" toggle in the settings or a button next to the input field.

    Option A (Toggle): A permanent setting to keep the reasoning steps visible alongside the final answer (perhaps in a collapsible sidebar or indented block).
    Option B (On-Demand): A "Reveal Logic" button that, when clicked after a response, expands the hidden reasoning steps used to generate that specific answer.
    Option C (Hybrid): Keep the current brief flash for casual users, but allow power users to "pin" the reasoning view for the duration of a session.
    

    Conclusion: This feature aligns perfectly with Proton's ethos of transparency and user control. By giving users the option to see the "engine room" of the AI, we can foster deeper trust and more effective collaboration. It transforms Lumo from a simple answer engine into a transparent thinking partner.

    Hello, Proton.

    Fair notice: I did allow Lumo to aid my crafting of this feedback and suggestion. I thank you folks for all that you do!

    Susana

    To: Lumo Product Team / UX Design From: Susana [User]

    Observation: Following the recent UI update (April 2026), I noticed that Lumo briefly displays a numbered list of internal reasoning steps (e.g., "User is asking X, analyzing data Y") before posting the final response. While this currently disappears quickly, it offers a unique glimpse into the AI's logic.

    Insight: Rather than viewing this as a transient artifact, I believe this "reasoning trace" could… more

    1 vote

    Vote Vote Vote Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Vote

    We’ll send you updates on this idea

    0 comments  ·  Changes to existing features  ·  Edit…  ·  Delete…  ·  Admin →
    How important is this to you?

    Not at all You must login first! Important You must login first! Critical You must login first!

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)

    Close

    Close

    Submit Rating