Leo, the AI assistant in Brave, now offers end-to-end encrypted chat with GLM 5.3 Flash!
Your browser encrypts each prompt before it leaves your device, so only an attested Trusted Execution Environment (TEE) can decrypt it.
It's industry-leading privacy you can verify.
Those ads are more tempting in 2026 but they’re still just scams, folks.
ALT A comic strip featuring a scene from a superhero movie with three characters standing together, followed by a large, armored character sitting. The top text discusses DDR5 RAM pricing increases. A middle caption over the characters reads, "Where did that bring you?" A humorous button with "Click here to download more RAM!" is on the armored character, who responds, "Back to me.
New research by Citizen Lab reveals zero-click spyware delivered through targeted ads. They call it "AdFection."
Just viewing an ad can be enough to infect a device.
One of Citizen Lab’s recommendations is to block ads and trackers to reduce exposure to targeted ads like these.
Brave blocks ads and trackers on websites by default, with no settings changes or extensions needed.
Due to the complexity of AdFection, individuals and organizations at higher risk for spying such as press or activists should take further measures to protect themselves.
Read Citizen Lab's report for more recommendations: citizenlab.ca/research/how-d…
A new episode of #TheBraveTechnologist is here!
On this week's episode, The Digital Chamber CEO Cody Carbone (@CodyCarboneDC) explains why regulatory clarity is vital to the long-term growth of the digital asset industry.
Other key topics:
- What the CLARITY Act does for developers and token issuers
- How personnel changes at the SEC and CFTC shifted the regulatory environment for digital assets
Watch the full discussion, hosted by @LukeMulks, here: brave.com/podcast/e142/
NEAR AI is now available to @brave's 128M+ monthly active users!
I’m incredibly excited about this because the browser is the interface for the web and it already has the context that makes agents useful - your tabs, searches, documents, work, shopping, research, payments, etc.
That context is also some of the most sensitive data you have so if AI is going to live in the browser, privacy can’t just be a policy you’re asked to trust. It needs to be something you can verify yourself if you want to.
This integration has been cooking for a long time, and I’ve been incredibly impressed by the security diligence Brave brought to the process.
Now, Brave users have the option to select NEAR AI's models so their prompts are encrypted in the browser and only decrypted inside an attested confidential environment so neither Brave nor NEAR AI can read them.
This is @near_ai's first integration at this scale, and I genuinely couldn’t think of a better partner!
Learn more here: near.ai/blog/near-ai-cloud-b…
Leo, the AI assistant in Brave, now offers end-to-end encrypted chat with GLM 5.3 Flash!
Your browser encrypts each prompt before it leaves your device, so only an attested Trusted Execution Environment (TEE) can decrypt it.
It's industry-leading privacy you can verify.
For years, my research has asked: how do you prove an AI system is private, instead of promising it? Today we shipped an answer in production.
Your chats now travel like a sealed letter: @brave carries them, and only a verified enclave (@nvidia-TEE run by @near_ai) can open them
Leo, the AI assistant in Brave, now offers end-to-end encrypted chat with GLM 5.3 Flash!
Your browser encrypts each prompt before it leaves your device, so only an attested Trusted Execution Environment (TEE) can decrypt it.
It's industry-leading privacy you can verify.
Most AI privacy comes down to a policy you have to trust.
With Leo on @near_ai, your prompt is encrypted in Brave and only opens inside an enclave that signs a proof of where it ran and which model answered.
We walk through how it works and how to set it up:
piped.video/Shxc59Rkg68
Leo, the AI assistant in Brave, now offers end-to-end encrypted chat with GLM 5.3 Flash!
Your browser encrypts each prompt before it leaves your device, so only an attested Trusted Execution Environment (TEE) can decrypt it.
It's industry-leading privacy you can verify.
Brave Leo now runs on private AI inference from NEAR AI Cloud, by @near_ai.
Pick GLM 5.3 Flash in Leo and your chat is end-to-end encrypted into a hardware enclave (TEE).
@brave can't read it. NEAR AI can't read it. The cloud host can't either. 🧵
Leo, the AI assistant in Brave, now offers end-to-end encrypted chat with GLM 5.3 Flash!
Your browser encrypts each prompt before it leaves your device, so only an attested Trusted Execution Environment (TEE) can decrypt it.
It's industry-leading privacy you can verify.
You can now chat in @brave Leo with end-to-end encryption and verifiable privacy, powered by @near_ai Cloud.
Browsers are a primary computing interface that handle all of our personal information, so it’s critical that AI is utilized in a user-owned way.
Leo, the AI assistant in Brave, now offers end-to-end encrypted chat with GLM 5.3 Flash!
Your browser encrypts each prompt before it leaves your device, so only an attested Trusted Execution Environment (TEE) can decrypt it.
It's industry-leading privacy you can verify.
Leo, the AI assistant in Brave, now offers end-to-end encrypted chat with GLM 5.3 Flash!
Your browser encrypts each prompt before it leaves your device, so only an attested Trusted Execution Environment (TEE) can decrypt it.
It's industry-leading privacy you can verify.
These privacy protections won't slow you down, though.
In our testing, a 19.9-second chat request spent 297 microseconds encrypting the prompt and 5.76 milliseconds decrypting the streamed reply. That's only 0.0015% and 0.03% of the total wait time.
End-to-end encrypted Brave AI chat is available today in all channels (Nightly, Beta, and Release).
To get started, select "GLM 5.3 Flash" from the model selection menu in Leo.
We'd love your feedback as we work on additional improvements.