Lumo vs ChatGPT: who reads your prompt, what each keeps, and price, with Morse alongside

The Morse teamPublished October 31, 2026Facts checked October 31, 202613 min read

In Lumo vs ChatGPT, Lumo wins on privacy defaults and ChatGPT on models and tools, but neither keeps its own server from reading your prompt. As of October 31, 2026, Proton's Lumo doesn't train on your chats, shows no ads, and stores your history with zero-access encryption, while ChatGPT trains on Free, Plus, and Pro chats unless you opt out, shows ads on Free and Go, and keeps your history on OpenAI's systems until you delete it. Choose Lumo if you want those defaults from a Swiss company, choose ChatGPT if you want OpenAI's largest models, voice, and Codex, and choose Morse if you want your own browser to verify the enclave that reads an E2EE prompt before it leaves your device.

On October 31, 2026, we read Proton's own Lumo vs ChatGPT page, its Lumo page, pricing, privacy, memory, export, and getting-started help articles, the Lumo privacy policy (last modified September 17, 2026), the Lumo terms (last modified August 19, 2026), the Lumo security model post, and the Lumo 2.0 announcement. From OpenAI we read its help pages on data controls, deleting chats, Temporary Chat, memory, ads, model training, consumer data, and signed-out use, its US and EU privacy policies, its security page, its usage policies, and its pricing page. We ran no tests. Both companies earn some credit. Proton's comparison page marks ChatGPT "Yes" on 10 of its 17 rows instead of marking it down across the board, and OpenAI's help pages state each 30-day window in plain words. Our Lumo AI review covers Lumo alone, and Morse vs ChatGPT covers ChatGPT alone.

Lumo vs ChatGPT vs Morse at a glance

As of October 31, 2026. Every Lumo and ChatGPT cell links to the company's own page.

Is Lumo more private than ChatGPT?

Yes on what each company keeps afterward, and no on who reads the prompt. Proton says that after Lumo answers, "the data is erased," that it keeps "no logs of what you ask, or what Lumo replies," and that it stores no "timestamps, IP addresses, or chat context" (Proton). OpenAI keeps saved chats until you delete them, and a limited number of authorized staff and "trusted service providers" may read content to investigate abuse, give support, handle legal matters, or improve models unless you've opted out. Its own advice is "Please do not enter sensitive information that you would not want reviewed or used" (OpenAI).

The step before that is the same on both. Proton's security post says "we need to provide the LLM with cleartext messages," and adds that "with each new message, we must send the full conversation from top to bottom, because the LLM server is memoryless" (Proton). So Proton's model server decrypts the whole conversation at every message, then, Proton says, forgets it. OpenAI encrypts your content "at rest and in transit between you and OpenAI" (OpenAI), which leaves OpenAI holding the keys. In both apps, what happens to a prompt after the server reads it is a policy you take on trust. The review explains Proton's user-to-Lumo encryption in more depth.

Does Proton's Lumo vs ChatGPT table hold up?

Mostly, yes. Proton's comparison page sets Lumo and ChatGPT side by side in 17 rows (Proton). We checked the rows about data against OpenAI's own pages, read the same day.

We left out the "Open-source models" row, which OpenAI's pages don't address, and eight feature rows where Proton marks both apps "Yes." Two of Proton's "No" cells say more than OpenAI's pages do. The ad row holds for Free and Go only, and the sharing row leaves out that OpenAI says it doesn't share conversations with advertisers or sell data.

The results for "lumo vs chatgpt" on October 31, 2026 were led by Proton's own pages and hands-on reviews of answer quality. Proton's table has no row for whether the server that answers can read your prompt, and on that row Lumo and ChatGPT would both get a "No."

Does ChatGPT train on your chats, and does Lumo?

ChatGPT does by default on Free, Plus, and Pro, and Lumo doesn't. Morse vs ChatGPT lists the exceptions to ChatGPT's opt-out, and OpenAI's data controls page, updated the day before we read it, adds one more. During Daybreak Trusted Access sessions, OpenAI "may use content from your interactions to improve our models even when 'Improve the model for everyone' is off" (OpenAI).

Proton says "Lumo doesn't use your chats to train the AI models" (Proton). Lumo has a feedback route too. Its feedback button sends "an anonymized version of your prompt and Lumo's response" to Lumo's team, and feedback on Apertus answers "may also be shared with ETH Zurich and EPFL" for purposes that include "training of the Apertus models" (Proton). Proton anonymizes what its feedback button sends, while a thumbs up or down on ChatGPT can put "the entire conversation" into training (OpenAI). The history of OpenAI's rules is in does ChatGPT train on your data.

Is Lumo's Ghost mode the same as ChatGPT's Temporary Chat?

Close, though OpenAI describes its version in more detail. A ChatGPT Temporary Chat stays out of your history, doesn't create or update memories, isn't used for training, and shows no ads, but "OpenAI may keep a copy of a temporary chat for up to 30 days for safety purposes" (OpenAI; OpenAI). Pick Unpersonalized before your first message, because a Personalized temporary chat "can use your existing memories, custom instructions, and plugins," and you can't change it once the chat starts.

Proton describes Ghost mode as a way to "keep Lumo focused on a single conversation by temporarily cutting its access to chat history," and says Lumo "never logs your data, whether or not you use Ghost mode" (Proton). The Lumo pages we read don't say more about it, such as whether a Ghost chat can be saved later. Without an account, a Lumo guest's conversation "is erased at the end of each session" (Proton), and a signed-out ChatGPT chat stays "only for the current session on the same browser" (OpenAI).

Where do Lumo and ChatGPT keep your history and memories?

ChatGPT keeps both on OpenAI's systems, "in the US and around the world" (OpenAI), and Morse vs ChatGPT lists each copy that outlives a deleted chat. Lumo keeps an encrypted copy on Proton's servers.

Proton's pages describe Lumo's history in two ways. Its privacy article says "the only record of the conversation is on your device," and a few lines later that chats are "stored on your device and synced to Proton servers using zero-access encryption" (Proton). Both hold, because the copy on Proton's servers is one Proton says it can't open. A Free account keeps chats "for 7 days only" (Proton), and Plus adds "extended history retention" (Proton). Lumo's Memory works on Free and paid plans with a Proton Account, and "memories are zero-access encrypted" (Proton). Proton says it discloses "the limited user data we possess" only on "a binding request coming from the competent authorities" (Proton). OpenAI says that when it receives "a lawful subpoena," it may keep data "for the duration of the relevant legal or regulatory obligation" (OpenAI).

Does ChatGPT have ads, and does Lumo?

ChatGPT does on Free and Go, and Lumo doesn't. OpenAI started testing ads in the US on February 9, 2026. They appear below a response, and OpenAI is also testing a multi-image ad format that shows during image generation (OpenAI). Every way to turn them off is in how to turn off ChatGPT ads.

Proton's page says "Lumo never shows ads and never shares your data with third parties" and links to OpenAI's ad announcement (Proton).

Which one refuses more?

Neither publishes a refusal rate, and we ran no test. Proton says it "has implemented safeguards in Lumo to prevent harmful or offensive responses" (Proton), and Lumo's terms bar illegal activity, CSAM, and harassment, and say output is no "substitute for professional advice" (Proton). OpenAI's usage policies, in a version effective November 20, 2026, go into more detail. They bar uses such as "unsolicited safety testing" and "circumventing our safeguards," along with AI tools that give others "tailored advice that requires a license" without a licensed professional (OpenAI).

Which models and features does each have?

Lumo runs open models such as Qwen 3.5 and GLM 5.3 on Proton's servers (Proton), with web search, image generation, Projects, and memory, all covered in our Lumo AI review.

ChatGPT runs OpenAI's own. Free gets GPT-6 Luna and GPT-5 Thinking Mini, Plus adds GPT-6 Sol, GPT-6 Astra, GPT-6.1 Sol, and legacy models, and Pro adds "Pro reasoning powered by GPT-6 Astra" (OpenAI). Every plan includes voice, deep research, and Codex, with lower limits on Free. Lumo has no GPT model, and ChatGPT has none of the open models Lumo runs.

Is Lumo cheaper than ChatGPT?

Yes. Lumo Plus costs $7.01 a month less than ChatGPT Plus, or $10.01 a month less if you pay Proton for a year.

Both put a ceiling on "unlimited." Lumo's terms define it as "no fixed numerical cap" on "ordinary, good-faith use," with throttling for use far above typical (Proton), and ChatGPT's unlimited text chats are "subject to abuse guardrails." Lumo also accepts "anonymous cash or Bitcoin payments" (Proton).

Where Lumo and ChatGPT are better

• Lumo, on jurisdiction and payment. It runs under Swiss law on servers Proton controls, and you can pay in cash or Bitcoin.
• Lumo, on Proton's apps. A free Lumo account also opens Proton's VPN, email, calendar, and cloud storage, Lumo reads files from Proton Drive, and it has iOS and Android apps (Proton).
• ChatGPT, on free use. Free has unlimited everyday text chats, where Lumo's limits are unnumbered and Morse gives 15 messages a day.
• ChatGPT, on tools. Voice, Codex, ChatGPT Work, and Pro's GPT-6 Astra reasoning have no match in Lumo, and Morse is a web app.

How Morse handles the server both ask you to trust

Every difference above is a rule about what a company does after its own server has read your prompt in the clear. Proton promises to forget it, and OpenAI keeps it under its 30-day windows, training switch, and ad signals. Morse is built so that, in E2EE, the only machine that can read your prompt is an enclave your own browser has checked, and in every mode your history stays off our servers.

• E2EE, for prompts you don't want any company's server to read. Before anything leaves your device, your browser checks the enclave's hardware attestation itself, against versions pinned in the app, and then encrypts the message so only that enclave can read it. Our servers relay ciphertext they can't read, and the reply is decrypted only in your browser. Open the lock on any E2EE reply to see each check that passed, or download the full report. The enclaves are run by third-party providers, not by Morse, and the guarantee depends on the enclave hardware and attestation working as designed. E2EE runs open-weight models such as GLM, Kimi, and DeepSeek, without web search or image generation, and comes with Pro at $19 a month.
• Anonymous, for the GPT models Lumo doesn't carry, without the provider knowing it's you. Our proxy removes your account, email, and IP address and sends the message under Morse's own account. The provider sees the words but not who sent them, and may keep them for a limited time to check for abuse, under its own policy. Leave your name and identifying details out of the prompt.
• Private, the default, for open-weight models such as GLM, the family Lumo also runs, on hosts that keep nothing. Morse has zero data retention agreements with its inference, search, and other providers. Where a provider confirms zero retention on each reply, Morse checks it and refuses a reply that doesn't; otherwise Morse relies on the provider's commitment.

In every mode, your history is stored only in your browser, encrypted, on Free as well as paid plans. Opening it takes two keys, one that never leaves your device and one our server gives you at sign-in. There's no training switch to find, because Morse doesn't train AI on your content, and no ad setting, because Morse shows no ads, sells no data, and runs no advertising trackers or third-party analytics. Temporary chats are kept in memory only. Because we don't store chats, we usually have very little to give authorities. Clearing your browser deletes your chats, and we can't recover them, so keep a copy of anything you need. The privacy modes page shows what anyone can see of a message in each mode.

Choose Lumo, ChatGPT, or Morse

• Choose Lumo if you already use Proton Mail or Drive, want Swiss law and servers Proton controls, want to pay in cash or Bitcoin, and are happy with open models.
• Choose ChatGPT if you want OpenAI's largest models, voice, Codex, and unlimited free text chats, and will turn off training and pick Unpersonalized temporary chats yourself.
• Choose Morse if you want your own browser to verify the enclave before a prompt leaves it; GPT, Claude, and Gemini without your account attached; open-weight models such as GLM, DeepSeek, and Kimi in the same place; and history kept encrypted on your device, with no ads or training on any plan.

Send a first message from the Morse home page without an account, or start free with no card. E2EE chats come with Pro, at $19 a month.

Morse publishes this comparison. We've tried to keep it fair; if something's wrong or out of date, email hello@trymorse.ai.

Proton and Lumo are trademarks of Proton AG. Morse is not affiliated with or endorsed by Proton AG.

ChatGPT is a trademark of OpenAI. Morse is not affiliated with or endorsed by OpenAI.