Morse vs Brave Leo: who reads your prompt, and who checks the enclave

The Morse teamPublished November 6, 2026Facts checked November 6, 202611 min read

In Morse vs Brave Leo, both keep your chat history on your device and neither trains on your chats, and they split on who can read a prompt and who checks. As of November 6, 2026, Brave runs 13 of Leo's models on infrastructure it says it operates itself, under a policy that it keeps nothing, and encrypts one model, GLM 5.3 Flash, end to end, with Brave's own server checking the enclave on your browser's behalf. Morse lets you choose per chat, and in an E2EE chat your browser checks the enclave itself before anything leaves your device. Choose Brave Leo if you want AI built into your browser, a model that runs on your own machine, or encrypted chat on a free plan; choose Morse if you want the check on your own device, GPT, Claude, and Gemini without your name on them, and over 30 models in one account.

On November 6, 2026, we read Brave's Leo page and its FAQ, the Leo section of Brave's browser privacy policy (dated September 23, 2026), five Leo articles in Brave's help center (models, Bring Your Own Model, ChatGPT, chat history, and memory), Brave's premium pricing page, and both of Brave's posts on Leo's trusted execution environments, from November 20, 2025 and October 15, 2026. We ran no tests. Brave earns credit for publishing the code for every stage of its encrypted pipeline, and its October post names its own remaining gap in plain words. For Leo next to Duck.ai and Lumo, see our three-way comparison.

Morse vs Brave Leo at a glance

As of November 6, 2026. Every Leo cell comes from a Brave page linked here or in the sections below.

Can Brave read your Leo chats?

For 13 of Leo's models, Brave's servers process your prompt, and Brave's no-logs policy is the protection; for GLM 5.3 Flash, Brave says it can't read it at all. Brave's help center says every Leo model runs on Brave's own infrastructure (Brave), and a June 30, 2025 post says that for Claude and DeepSeek this means AWS Bedrock "with the same zero-retention configuration," which ended "the 30-day data retention that previously applied to Claude models under Anthropic's standard API terms" (Brave). Our three-way comparison quotes the rest of the retention policy.

Brave is candid about what that policy is. Its October post describes the protections its first enclave release ran under, "encrypted in transit, never logged, and discarded once a response was generated," and then says: "Those protections rested on our policies, not yet on cryptography" (Brave). Those are the same protections every Leo model except GLM 5.3 Flash runs under today.

Is Brave Leo end-to-end encrypted?

One model is: since October 15, 2026, GLM 5.3 Flash chats are sealed in your browser for a NEAR AI enclave, and Brave's gateway checks that enclave before your browser gets its key (Brave). Our three-way comparison walks through that relay and verifier, and what Oblivious HTTP is covers the protocol.

Brave's post describes its check closely enough to set beside Morse's.

Brave's Leo column comes from its October 15, 2026 post (Brave). Morse's guarantee rests on the enclave hardware and attestation working as designed, and our guide to verifying an AI enclave's attestation explains what each kind of check proves.

Every route a Leo prompt can take, and Morse's nearest

Leo's privacy depends on what you pick, and Brave's pages describe at least six routes. Each Leo cell comes from the Brave page linked in its row, read November 6, 2026; the Morse column shows the mode you'd use for the same job.

Brave's help center has an article on how to "add my ChatGPT account to Brave Leo," which takes the fifth route, through an OpenAI API key (Brave), and Brave's Bring Your Own Model article warns that the third party may "Use your data for model training" (Brave). On Morse, the same models go through Anonymous mode under Morse's own account, and the provider may keep a message for a limited time to check for abuse, under its own policy. A Leo chat in a Private Window or Tor window is "never saved to history" (Brave); Morse's temporary chats are kept in memory only.

Where does Leo keep your chat history and memory?

On your device. Brave's help center says Leo conversations are "stored locally on your device only," "encrypted when stored," and removable under Settings, then Leo, with Clear Leo data, or by turning off Store my conversation history (Brave). Turn on Sync and the history joins your other Brave data in an encrypted file on AWS, with "the decryption key, which is stored on your devices," and an unused sync chain expires after 12 months (Brave).

Leo's memory is local too, and you add each memory yourself. Brave notes that memories "are included with your prompts," so they travel to the model with each message (Brave). Morse also keeps history in your browser, encrypted, and locks it when you sign out. Opening it takes one key that never leaves the device and one Morse's server gives you at sign-in, so a copy of the browser data alone can't be read.

Which models does each offer?

Leo has 5 models for free users and 13 on Premium in Brave's help center table, which our three-way comparison lists in full (Brave), plus any model you connect through Bring Your Own Model. That table, read November 6, 2026, leaves out GLM 5.3 Flash, the one encrypted model, which Brave's October post says is "available today to every Brave user" (Brave), and the Leo page's FAQ still names Mixtral and Llama, which the table doesn't list (Brave). Open Leo's model menu to see what you'll get.

Morse has over 30 models, open and closed, in one account, from GPT, Claude, and Gemini to DeepSeek, Kimi, GLM, and Grok, with Characters for roleplay and companion chats and Studio for images. Paid plans add unfiltered models, for dark fiction and straight medical and legal answers that other apps refuse; none of the Leo pages we read describes an unfiltered option.

Is Leo Premium cheaper than Morse Pro?

Yes. Leo Premium is $14.99 a month or $149.99 a year, with a 7-day free trial (Brave). Morse Pro is $19 a month, or $17 billed yearly, and Max is $49. Free Leo has "Reasonable rate limits" with no number published, while Morse Free gives you 15 messages a day, 3 images a day, and web search, with no card.

The plans buy different things. Premium adds models and "Higher rate limits" (Brave). Morse Pro adds E2EE chats, unlimited Pro models such as Claude Sonnet 5.5, DeepSeek V4.1 Flash, and Grok 4.7, unfiltered models, memory across chats, and 250 credits a month for frontier models.

Where Brave Leo is better

• A model on your own machine. Bring Your Own Model runs a local model on desktop, and Brave says the prompts "never leave your machine."
• Encrypted chat that can search. GLM 5.3 Flash calls Brave Search from inside the enclave. Morse's E2EE chats use only tools that run on your device, so they skip web search.
• Encryption on the free tier. Brave says its encrypted model is open to every Brave user; Morse's E2EE chats come with Pro.
• Claude under zero retention. Brave's June 2025 post says Claude runs through AWS Bedrock in a zero-retention configuration. On Morse, Claude runs in Anonymous mode, where the provider sees the message without knowing it's you and may keep it for a limited time to check for abuse.
• The page you're on. Leo summarizes the tab you're reading, PDFs, and Google Docs, and can organize your tabs (Brave).
• Open code for the encrypted path. Brave publishes its client, gateway, and verifier code, and NEAR AI publishes what runs in the enclave (Brave).
• A lower price. Premium is $14.99 a month, against $19 for Morse Pro.
• History on every device. Brave Sync carries Leo history between your devices, encrypted with a key kept on them.

How Morse puts the enclave check on your device

Leo's privacy rests on Brave twice: on its no-logs policy for 13 models, and on its gateway's word that the enclave is genuine for the fourteenth. Morse is built so the step you'd most want to check runs where you can see it, in your own browser.

• E2EE, for a prompt no server should read or vouch for. Your browser checks the enclave's hardware attestation against versions pinned in the Morse app, and fails the check if, say, the OS image hash differs from the pinned one. Then it encrypts the message so only that enclave can read it, Morse's servers relay ciphertext they can't read, and the reply is decrypted only in your browser. Each E2EE reply has a lock you can open to see every check that passed, and a full report you can download. The enclaves are run by third-party providers, not by Morse, and the guarantee rests on the enclave hardware and attestation working as designed. E2EE runs open-weight models like GLM, Kimi, and DeepSeek, without web search or image generation, and comes with Pro.
• Anonymous, for GPT, Claude, and Gemini without your name. Our proxy removes your account, email, and IP address and sends the message under Morse's own account, so the provider sees it without knowing it's you. It may keep it for a limited time to check for abuse, under its own policy, so leave out names and details you'd rather keep.
• Private, for DeepSeek, Kimi, GLM, and Grok 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.

In every mode, history stays in your browser, encrypted. Clearing the browser deletes it and Morse can't recover it, so keep a copy of anything you need. We don't train AI on your content, sell your data, or show ads, and because we don't store chats, we usually have very little to give authorities (privacy policy).

Brave Leo or Morse: which to pick

• Choose Brave Leo if you want Claude in the zero-retention setup Brave describes, encryption without paying, and Leo history carried between your devices by Brave Sync.
• Choose Morse if you want the enclave checked by your own browser against pinned versions, a report of that check you can download, Anthropic's and OpenAI's models with your account and IP address removed, unfiltered models on paid plans, and history that locks when you sign out.

Try a chat from the home page without an account, or start free with 15 messages a day and no card. With Pro, the lock on each E2EE reply lists the checks your browser ran.

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

Brave and Leo are trademarks of Brave Software, Inc. Morse is not affiliated with or endorsed by Brave Software, Inc.