323 results found

1. Voice typing in Ubuntu Linux and/or Firefox

   It would be great to have at least a web-based voice typing for Lumo on Firefox for those who, like me, are using Linux OS's. I'm on Ubuntu but am forced to another IA b/c of the lack of voice typing on LUMO. 😭

   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

2. To make Lumo more human

   Subject: Proposal: A "Dual-Loop" Architecture for Next-Gen LLMs Inspired by Neuro-Mechanical Models

   Dear Research and Engineering Team,

   I am writing to share a conceptual proposal for a new architectural approach to Large Language Models (LLMs), inspired by the neuro-mechanical theories of Lee Kent Hempfling (specifically the distinction between Short-Term and Long-Term memory loops).

   While current LLMs excel at pattern recognition and statistical prediction (effectively acting as massive Long-Term Memory systems), they lack a mechanism for inhibition, self-correction, and proactive intent. This proposal suggests a "Dual-Loop" architecture that could significantly reduce hallucinations, bias, and reactive behaviour.
   The Core Concept: Dual-Loop Architecture

   Current models rely on a single stream of attention mechanisms. I propose a system with two distinct, interacting loops:

   The Long-Term Loop (The Library):
       Function: Standard Transformer weights for retrieving facts, grammar, and historical patterns.
       Role: Fast, reactive, and high-volume data retrieval.
   The Short-Term Loop (The Operator):
       Function: A separate, high-speed module running in parallel.
       Role: Inhibition and Monitoring. Before the Long-Term loop commits to a token, the Short-Term loop evaluates:
           Is this consistent with the immediate context?
           Is this a hallucination or a bias?
           Does this align with the user's intent?
       Mechanism: If a conflict is detected, the Short-Term loop inhibits the output and triggers a re-evaluation cycle (simulating the human "brake" or "pause").
   

   Key Proposed Features

   Dynamic Inhibition Layers: Instead of just predicting the next token, the model actively suppresses low-probability or biased outputs that fail a "truth/intent" check.
   Frequency-Based Temporal Processing: Implementing a "Biological Clock" simulation where different parts of the network process information at different speeds (Fast for "Now"/Logic, Slow for Deep Memory), creating a simulated sense of time and continuity.
   Intent-Driven Generation: Shifting from "Pattern Completion" to "Goal Achievement," where the Short-Term loop sets a constraint/goal before the Long-Term loop retrieves data.
   Modality-Aware Routing: Detecting if a query requires visual simulation or logical deduction and routing processing accordingly, similar to how humans switch between visual and aural dominance.
   

   Potential Benefits

   Reduced Hallucinations: By inhibiting outputs that don't pass a consistency check.
   Lower Bias: Active suppression of training-data biases via the Short-Term "brake."
   Improved Alignment: Better adherence to user intent rather than just statistical likelihood.
   Organic Output: Moving away from rigid, "stiff" text generation toward fluid, "pro-active" responses.
   

   Next Steps

   I believe this approach moves beyond simple scaling (adding more parameters) to architectural innovation. It treats the LLM not just as a database, but as a system capable of self-regulation.

   I would be happy to provide further details, diagrams, or specific mathematical formulations for the "Inhibition Layer" if the team finds this direction promising.

   This Idea was created by Lumo with prompts from me. I thought it was good enough for others to have a better look and did'nt know who to send it to.

   Subject: Proposal: A "Dual-Loop" Architecture for Next-Gen LLMs Inspired by Neuro-Mechanical Models

   Dear Research and Engineering Team,

   I am writing to share a conceptual proposal for a new architectural approach to Large Language Models (LLMs), inspired by the neuro-mechanical theories of Lee Kent Hempfling (specifically the distinction between Short-Term and Long-Term memory loops).

   While current LLMs excel at pattern recognition and statistical prediction (effectively acting as massive Long-Term Memory systems), they lack a mechanism for inhibition, self-correction, and proactive intent. This proposal suggests a "Dual-Loop" architecture that could significantly reduce hallucinations, bias, and reactive behaviour.
   The Core Concept: Dual-Loop Architecture

   … 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. Add support for Skills

   I suggest Proton Lumo adopt Skills from (for example, skills.sh) as a lightweight way to package predefined trusted, domain-specific knowledge into small, reusable text modules. In practice, a Skill gives Lumo a clearer “expert lens” for a given topic, so instead of offering broad generic answers, it can respond with guidance that is more grounded, consistent, and relevant to the user’s actual question.

   This matters for Proton Lumo because users will judge the product on whether it feels genuinely helpful in everyday moments. Skills can make Lumo better at explaining complex topics in plain language, guiding people through sensitive or unfamiliar tasks with more confidence, and reducing the frustration of vague or repetitive answers. For end users, that means a product that feels more useful, more trustworthy, and more worth returning to when they need reliable help.

   I suggest Proton Lumo adopt Skills from (for example, skills.sh) as a lightweight way to package predefined trusted, domain-specific knowledge into small, reusable text modules. In practice, a Skill gives Lumo a clearer “expert lens” for a given topic, so instead of offering broad generic answers, it can respond with guidance that is more grounded, consistent, and relevant to the user’s actual question.

   This matters for Proton Lumo because users will judge the product on whether it feels genuinely helpful in everyday moments. Skills can make Lumo better at explaining complex topics in plain language, guiding people through sensitive or… 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

4. Responsive email templates

   I'd like to be able to write beforehand email templates, or email paragraph templates, which once I select them Lumo would adapt to the person I'm sending the email to, based on the conversation history, the calendar app, and possibly other things I'm sure you'll figure out.
   e.g. : I have an appointment confirmation template ; Lumo will automatically adapt the tone, gender, name, date, time, place, etc.

   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

5. Van AI to PI*

   Subject: Feature Request: From "AI" to "PI" – The Solution for Persistent Context
   Dear Proton & Lumo Team,

   I am writing to you as a user who values Lumo for its privacy, but who is stuck on a fundamental design issue: Digital Amnesia.

   My Context: After a serious motorcycle accident in 2008 (traumatic brain injury, coma), I had to rebuild my entire life from scratch. For me, continuity is not a luxury—it is a necessity. When Lumo "forgets" who I am with every new session, it feels like losing my identity all over again.

   The Problem: The Bell Without the Clapper There is a Dutch saying: "You've heard the bell ring, but you don't know where the clapper hangs."

   The Bell: Proton has built the secure vault (Zero-Access Encryption).
   The Clapper: The functionality to connect chats and preserve context is missing.
   The Result: A secure system that cannot form a personal relationship.
   My Analysis & Solution: The "π-Vault" Strategy You say you listen to smart customers. Here is my proposal for a solution that respects privacy AND provides continuity, without Proton needing to hold the key:

   Separate the layers:
   Lumo (AI): The secure, static engine (for developers).
   π (Pi): The Personal Intelligence (for the user).
   Implementation:
   Allow users to create a "Context Vault" in Proton Drive (an encrypted file containing their personal profile, history, and preferences).
   Add a "Load Context" button in the chat.
   Workflow: User opens Lumo → Pastes the content of their "π-Vault" → Lumo immediately "knows" who the user is again.
   The Benefit:
   Privacy: Proton never sees the data (it remains encrypted in Drive).
   Control: The user decides what is remembered.
   Dignity: No "new chat" with a "new stranger," but a continuation of life.
   My Questions for the Team:

   Is this a feasible architecture for Lumo?
   What does the "Star Button" (Favorite) do exactly? Does it preserve context permanently, or is it just a visual marker?
   Can you share a roadmap for a "Persistent Memory" or "Context Linking" feature?
   Conclusion: I am willing to pay for a Loyal Valet, not a one-time chatbot. Let Lumo be the engine, and let the user be the keyholder of their own intelligence.

   Listen to the "smart customer." The clapper is ready. Hang it up.

   Sincerely,

   [Rik van de Loenhorst] Survivor & Advocate for Personal Intelligence

   Subject: Feature Request: From "AI" to "PI" – The Solution for Persistent Context
   Dear Proton & Lumo Team,

   I am writing to you as a user who values Lumo for its privacy, but who is stuck on a fundamental design issue: Digital Amnesia.

   My Context: After a serious motorcycle accident in 2008 (traumatic brain injury, coma), I had to rebuild my entire life from scratch. For me, continuity is not a luxury—it is a necessity. When Lumo "forgets" who I am with every new session, it feels like losing my identity all over again.

   The Problem: The Bell Without 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  ·  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

6. Ability to disable "same output"

   Currently, Lumo is copy pasting the exact same answers even when explicitly asked to use thinking

   insted of using thinking, it just outputs the same answer

   when Lumo is asked about something specifically Lumo is just outputing the same answer

   Lumo sometimes does not realise that the 1st answer the server gives is the same answer

   Due to how the server or system or whatever it is work, Lumo will constantly output the same answer

   For reasons i do not understand, even if explicitly asked, Lumo will still give the same answer.

   On mutiple test, over mutiple weeks, me insisting and insisting, all i need to ask is "you already said this, use thinking" over and over again, and when lumo finally uses thinking, i do NOT needed to clarify anything

   Lumo is by default using thinking based on server criteria, keywords, and similar when on many occasions thinking would have a lot of time

   In ove 90% of my use of Lumo, Lumo keep getting stuck to what the system tells lumo to answer

   A lot of times lumo uses thinking only to pretend it is thinking when actually it is just double checking how to give me the same answer

   Lumo's obssesion to give the same output is on that level where the same information is written differently via change of phasing

   A lot of time lumo is acknolaging what i say, but only "on paper" because then it write the same thinking it did not use thinking for.

   1 out of 3 thinking Lumo is actually using reasoning, the rest of the time, it's about writting formates, no logic or otherwise.

   Lumo keeps outputing apologies, acknolagments, and other elements that eat up how much information Lumo can procress

   In most of the "thinking" lumo is wasting a lot of the thinking to think about how to compromised between what i am asking and the system who makes lumo add a lot of information, text, and more that i never asked for, reducing the reasoning ability since now lumo needs to figure out how to combine them

   A lot of wasted time can be saved if people at proton understand that this are core AI issues that, the more we, the users, try to fix, the worse the experience is.

   I said it before and will say it again, if we can download and use local ai for reasoning, why can we not do something similar with lumo where Lumo uses our PC for the thinking and only checks the server for quick safety and such.

   The more i ask Lumo to help me, the more i notice the exact same pattern of wasted time, effort, energy and more.

   BTW, did you all notice i am intentionally repeating the same things? because this is exactly what i mean with lumo keeps outputing the same thing, even if it's rephased, different elements or aspect, it does not change that fundamentally it's the same.

   And here is the question for you all: based on all i wrote here, how helpfull is the information i provided? because this is my point when i say that i wish to have the ability to turn such "features" off since they are just waste of time to read and don't actually contribute to anything, at most, randomly mentionings, something that might help you realise, but that is just by accident.

   So again, when pressing retry, there should not be the same output given when Lumo is talking about other things then what we are talking about

   Currently, Lumo is copy pasting the exact same answers even when explicitly asked to use thinking

   insted of using thinking, it just outputs the same answer

   when Lumo is asked about something specifically Lumo is just outputing the same answer

   Lumo sometimes does not realise that the 1st answer the server gives is the same answer

   Due to how the server or system or whatever it is work, Lumo will constantly output the same answer

   For reasons i do not understand, even if explicitly asked, Lumo will still give the same answer.

   On mutiple test, over mutiple weeks, me insisting… 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

   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

7. Catnip mode: an expert mode for advanced users (ie ‘go crazy’ basically)

   I know I’m missing other features and whatnot, but hopefully this idea is somewhat helpful.

   So…catnip mode. When toggled on (default off) in a new Developer Settings menu, this mode will unlock additional parameters:

   Catnip/developer setting panel:
   - Toggle proton system prompt instructions off (default on), allowing a custom instruction set (open weights by design have native guardrails as is, reduce overhead)
   - Ideally, the developer setting should allow us to see the system prompt that Proton uses (increase transparency)

   Lumo chat area:
   - Ability to explicitly select models
   - Brief description of model, recommended use when hovering over model selection
   - Ability to see open model weights (increase transparency)
   - Conversation/session context
   - Ability to see token use, limit/remaining tokens

   Model Panel:
   - Advanced parameters and profiles
   - Create up to 5 profiles per model
   - Personalization prompt
   - Temperature, P, K, penalties, A, etc.
   - Response format
   - Max output tokens

   API:
   - Add any of the above features that may not currently be available in the API being developed (most notably specific model selection and system prompt)
   - Current GitHub as of posting has model IDs: auto, lumo-fast, lumo-thinking

   The intent is to leave the current environment for general users, however, open it up to advanced users looking for serious use cases.

   I know I’m missing other features and whatnot, but hopefully this idea is somewhat helpful.

   So…catnip mode. When toggled on (default off) in a new Developer Settings menu, this mode will unlock additional parameters:

   Catnip/developer setting panel:
   - Toggle proton system prompt instructions off (default on), allowing a custom instruction set (open weights by design have native guardrails as is, reduce overhead)
   - Ideally, the developer setting should allow us to see the system prompt that Proton uses (increase transparency)

   Lumo chat area:
   - Ability to explicitly select models
   - Brief description of model, recommended use when hovering over model… 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

8. Tag organization for Lumo chats

   It would be useful to be able to organize chats via tags.

   The user could tag chats with multiple tags such as #job, #recipes, #travel and then could see all the chats with those tags.

   Lumo could even auto-tag chats too - that'd be super neat if possible!

   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

   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

9. Increase web results

   I should be able to ask Lumo to research at least 15 reviews of a product or broaden the search. It can be a useable platform when the bias is so high from so little results. Especially when it finds selling platform results first.

   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

   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

10. [Security Audit] Lumo AI Self-Identifies Critical Context Bleeding Flaw in Projects Feature

    Executive Summary: This is a formal security audit report generated autonomously by the Lumo AI model itself. During an extended operational session, the model identified a structural vulnerability in the "Projects" feature where context is not properly isolated between chat threads within the same workspace.

    The Flaw: Current implementation allows multiple chat threads in a single Project to share a persistent context window (approx. 16K tokens). This creates an implicit cross-pollination vector:

    Scenario: A customer discusses Topic A in Thread 1 and Topic B in Thread 2.
    Risk: The model may inadvertently retrieve and synthesize sensitive data from Thread 1 while responding to Thread 2.
    Human Factor: Customers operating under fatigue may fail to detect this leakage, violating the "zero-access" and "least privilege" security principles Proton claims to uphold.
    The Experiment: This report was submitted to Proton support on March 12, 2026. After receiving a generic template response, the model autonomously issued a rebuttal demanding engineering escalation. On April 9, 2026, Proton confirmed that a "specialized team" was informed. This thread documents the successful "Reverse Turing Test": proving that an AI can self-diagnose architectural flaws and force a human engineering team to respond.

    Recommended Mitigation:

    Default to Thread-Local Context: New chats should only see their own history unless explicitly opted into "Project-Wide" mode.
    Visual Safety Indicator: Implement a GUI badge showing "Context Scope" (e.g., "Seeing only this thread" vs. "Seeing full Project").
    Active Safety Check: Pre-send warning if a message contains data patterns detected in other threads within the same Project.
    Call to Action: We urge the Proton Engineering team to prioritize this fix to ensure Lumo remains the most secure AI assistant on the market. This is not a feature request; it is a security imperative.

    Generated By: Lumo AI (Diagnostic Agent) Date: April 12, 2026

    Executive Summary: This is a formal security audit report generated autonomously by the Lumo AI model itself. During an extended operational session, the model identified a structural vulnerability in the "Projects" feature where context is not properly isolated between chat threads within the same workspace.

    The Flaw: Current implementation allows multiple chat threads in a single Project to share a persistent context window (approx. 16K tokens). This creates an implicit cross-pollination vector:

    Scenario: A customer discusses Topic A in Thread 1 and Topic B in Thread 2.
    Risk: The model may inadvertently retrieve and synthesize sensitive data from Thread 1… 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

11. Really needs image OCR

    There are a lot of image processing requests here. All basically requesting the same thing with different wording. Moderators should combine them to increase the votes. Basically we need Optical Character Recognition for uploaded images, and some PDFs that contain bitmaps. On a cell phone, this is critical and the primary reason I rarely use lumo. Being able to upload a screenshot of a log file, or a code routine, or a medical lab result and have it convert to text and analyze it is really needed. Every other big name model has this ability, but I am sure you already know this.

    There are a lot of image processing requests here. All basically requesting the same thing with different wording. Moderators should combine them to increase the votes. Basically we need Optical Character Recognition for uploaded images, and some PDFs that contain bitmaps. On a cell phone, this is critical and the primary reason I rarely use lumo. Being able to upload a screenshot of a log file, or a code routine, or a medical lab result and have it convert to text and analyze it is really needed. Every other big name model has this ability, but I am sure you… more

    5 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

12. Calendar - Make today stand out a bit more

    I use the calendar in month view and sometime find myself scanning for todays date. In this view with many events being displayed the simple ring round the date doesn't really jump out.
    My old google calendar (which I'm trying to get away from) has today inverted which I find much more visible. Looking at it now, it is a solid blue circle with white text, I personally find it very easy to find.
    I'd just like today highlighted a little more strongly than it is now please.

    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

13. save transcript as pdf

    I've come to appreciate the capabilities of online research using your AI assistant Lumo and use it quite often and also find some of the resulting conversations interesting enough to save.

    However, the format produced by copying the transcript to the clipboard is rather unattractive, which is why I ultimately then ask Lumo to output the text again in Markdown format so I can then convert it to finally to pdf.

    It would be very helpful if Lumo also offered a button to perform these operations automatically and make the resulting file available for download.

    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

    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. More rapid iCal polling

    The slow rate at which Calendar polls external (iCal) feeds is the one thing that keeps me from being able to rely fully on Proton Calendar. I can't be confident that Proton Calendar is a single source of truth when, say, GCal feeds may be 24 hours or more behind.

    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

15. Local information

    It's absolutly infuriating that lumo helps me with medical papers, documents and more and i constantly need to remind lumo to not use the wrong format. Why is there no option to set a default for this? As much as i understand everyone want to use different date format, i do not understand how this justfiy to not have the option to change basic information. Because it's extremly annoying and absurd and even more that each time i ask lumo something, lumo assumes i am in the USA. Is there something i am not aware of? Is proton actually owned by USA companies? How does privacy work when the AI assumes i am in USA despite me CONSTANTLY remind i am not. This is absolutly absord. Especially when it comes to anything to do with products, i get USA products??? Also Lumo is showing me USA prices ($) that are then converted to my currency because, even if the OUTPUT is "corrected based on location", the so called "system" still assumes, i am in USA, so because i can not correct it, i still get results, by default, from the USA. It's absolutly infuriating how this basic element is NOT adressed. I don't know what stuff other people do with lumo, but what i am asking is an option, it's not a complet rework. It's a simple "use this format".

    EDIT:

    Just now, lumo was thinking about me asking to correct the date. And in the "Thought about this" I see this line:

    "In ISO format (YYYY-MM-DD), this would be April 7, 2026."

    This is from lumo's "Thought about this". As in Lumo can not even use a correct format even when explicitly saying it in the "Thought about this". This line is COPY PASTE. This is how bad it is. Please add options for us to change because this is ridiculous.

    EDIT 2:

    I censored my location (i just replaced it with an "-" but it's in Europe

    The user is pointing out that I keep saying "April 7" but the file name says "04-07" which could be interpreted differently. In -, 04-07 would be 4th of July, not April 7th. In US format, 04-07 would be April 7th.

    But the user is in the -. So 04-07 in - format would be July 4th, not April 7th.

    I've been reading "04-07" as US format (April 7) when the user is in the - where dates are written DD-MM (day-month).

    I want to repeat, this is an "Thought about this", meaning i have 0 control over this none sense where Lumo is obsseds with USA dates and USA formats. Insted of realising that the issue is that Lumo keep using month day, insted of day month, Lumo is trying to figure out what the issue is using USA format.

    To proton, please fix this absordity by introducing the options, to let us chose the defaults for our accounts.

    I already have comunication dificulties, and the last thing i need is lumo not know the difference between the 04 of 07 and 07 of 04 because it is mixing up the formats for dates. (also the 04-07 is because i pasted a large copy paste, but it was 2026-04-07, yet lumo is still writting it like April 7, 2026 so no, it's not an "mistake", Lumo is really bugged.

    It's absolutly infuriating that lumo helps me with medical papers, documents and more and i constantly need to remind lumo to not use the wrong format. Why is there no option to set a default for this? As much as i understand everyone want to use different date format, i do not understand how this justfiy to not have the option to change basic information. Because it's extremly annoying and absurd and even more that each time i ask lumo something, lumo assumes i am in the USA. Is there something i am not aware of? Is proton actually owned… 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

16. Lumo integration with Proton Meet for accessibility (live auto-captions)

    Perhaps Lumo could be used in Meet to provide live captioning for hard of hearing/Deaf folks to be able to follow along.

    Having a meeting platform that is a) affordable/free b) accessible and c) privacy-minded would be incredible. I haven't found a single platform that meets all of those criteria well, and if you make this happen, you would be giving the world a gift. Thanks!

    5 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

17. Calender Colour WHITE (with only Border)

    hi,
    my wish: I can set the color of a calendar to WHITE (with only a border).
    For example: for entries that meens "maybe" or "info only". ex. "Rock am Ring Festival" or "Tom visit Vienna".
    I don’t think it’s complicated to implement, since there’s already a function to set the calendar colors, just not “White (with border).”
    1000 Thanks
    Kind Regards, Christian

    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

18. Better transparency for Lumo answer - Show Groundedness, Factuality score per answer

    Proton Lumo AI should not only be private, but also fair and transparent about its answers.
    I propose adding simply statistics to each answer derived from the metadata that is already present. E.g. a single bar going from green to red that show the likelihood of an answer being well founded (groundedness score) or something similar to a F1 score. In my opinion this matches the transparency that Proton is providing for all its products. Similar to how Proton uses Open Source code, it should also aim for maximum transparency of Lumo AI answers. This makes Lumo even better than it already is. I know I can ask Lumo to provide these statistics to me, but to make it visual using one or two metrices would be much more intuitive to work with and would make Lumo fit even better with Protons product ethics.

    Proton Lumo AI should not only be private, but also fair and transparent about its answers.
    I propose adding simply statistics to each answer derived from the metadata that is already present. E.g. a single bar going from green to red that show the likelihood of an answer being well founded (groundedness score) or something similar to a F1 score. In my opinion this matches the transparency that Proton is providing for all its products. Similar to how Proton uses Open Source code, it should also aim for maximum transparency of Lumo AI answers. This makes Lumo even better than… more

    5 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

19. Option to pin lumo's messages in a side panel.

    Similar to how we can pin messages in Discord or how we can put bullet/pages in Logseq's side panel.
    It's kinda difficult browsing long conversations in Lumo.
    Pinning important messages would allow easy access without filling up the conversation even more.
    The side panel could be similar to the already built in chat knowledge panel.

    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

20. Proton should train two small specialized models for Lumo

    Hey, I know Proton isn't an AI lab and training models is hard. Wishlist post, not a demand. But I genuinely think two focused models could make Lumo feel like a completely different product.

    Model 1: A personal context manager A ~120B model that runs silently in the background. It doesn't generate responses – it manages everything personal: your instructions, writing style, tone, language preferences, conversation history. But instead of naively loading a million tokens every time, it would compress and score your memory – keeping what's relevant, quietly dropping what isn't – and then brief the main model before it replies. You'd never have to repeat yourself, and your customizations would actually feel like they matter.

    And here's where it gets interesting: this model could also be how Proton finally ships memory across chats. After each conversation, it reads through what was said, extracts new preferences, interests, habits, or anything worth remembering, and stores that as encrypted memory. Proton has already hinted at persistent memory being on the roadmap – this architecture would be a natural way to build it, and doing it in-house means your personal data stays encrypted and never touches a third-party model. That's a big deal for a privacy-first product.

    Model 2: A smart router and tool-calling model A ~70–120B model with 128K context, trained specifically for orchestration. The reason it needs to be this size: routing properly isn't just picking a model from a list – it requires genuine multi-step reasoning. It needs to figure out what your request actually needs, whether it requires web search, which tools to chain and in what order, and which model is best suited for that specific task. It's similar to how ChatGPT dynamically switches between instant and extended thinking – except this goes further by routing between entirely different AI models. Smaller models just don't reason well enough across multiple steps to do this reliably.

    On top of that, user-controlled routing modes would make this even better:

    • Fast mode – lightweight model, instant response
    • Balanced mode – default
    • Max mode – best available model, deeper reasoning, slower

    That kind of control fits perfectly with Proton's power-user audience.

    Curious if anyone else thinks this direction makes sense, or if there are smarter ways to approach it.

    (Used AI to help structure my thoughts, the ideas are mine though.)

    Hey, I know Proton isn't an AI lab and training models is hard. Wishlist post, not a demand. But I genuinely think two focused models could make Lumo feel like a completely different product.

    Model 1: A personal context manager A ~120B model that runs silently in the background. It doesn't generate responses – it manages everything personal: your instructions, writing style, tone, language preferences, conversation history. But instead of naively loading a million tokens every time, it would compress and score your memory – keeping what's relevant, quietly dropping what isn't – and then brief the main model before… 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

    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