#OPG
You know what I finally did? So I tried @OpenGradient Chat today.
Honestly, I'm tired of AI assistants that say we value your privacy then probably sell my data. Open Gradient does something different. Your messages get encrypted on your device, your identity stripped before hitting any model. Privacy enforced by cryptography and hardware, not a promise.
The positive? I tested Claude Fable 5 , one of the first platforms to integrate it. Works smooth. Also tried the Nous Hermes model in Private Chat. Uncensored. You can literally ask anything. No judgment, no logging.
Now the bugs I faced. First, switching between Gemini and Claude took about 8 seconds to load the model each time. Second, the image generation via Image Studio froze once at 95% for nearly 30 seconds before completing. Third, the credit balance display showed zero even after I purchased $5 worth, had to refresh the page twice. Fourth, one chat message got stuck sending for 20 seconds before erroring out, but resending worked fine. Nothing major. The core private chat worked. I even gave it to my friend to test too, and he ran into similar small delays but agreed the privacy setup feels legit.
Here's the deeper positive. Even with those small glitches, I never felt my data was exposed. The TEE attestation proof is real. And users who buy credits and use the chat regularly are eligible for the S2 OPG airdrop. That's a nice bonus for early testers.
So yeah, this is how AI should be. Private by default, verifiable by design.
So go try it guys, your secrets stay yours.
