OpenRouter Now Routes to NEAR AI (TEE-Backed, Zero Retention)
Nothing is loaded from YouTube until you press play, and then without tracking cookies.
About this video
NEAR AI is now a provider on OpenRouter. If you already route through OpenRouter, you can pin private inference for a model in a couple of clicks, with the queries decrypted only inside a TEE and no retention on the provider side.
Quick walkthrough of where to find it (search by model, not provider), how to pin NEAR AI as the endpoint, the code snippet OpenRouter hands you, and what the privacy panel shows for training, retention, and moderation. Also a note on keeping a fallback provider in your routing.
Scope note: private inference means the operator can't read your prompts while they're processed. Moderation and anything you do with the response afterward are still on you as the developer.
Chapters 0:00 NEAR AI is on OpenRouter 0:08 Finding it in the model list 0:17 Pinning NEAR AI as the provider 0:41 Grabbing the code snippet, and why to keep a fallback 0:56 Throughput and uptime 1:04 The privacy panel: no training, zero retention 1:22 Wrap
Transcript
315 wordsCaptions as YouTube publishes them, usually machine-generated, so product and people names can be misheard.
- 0:00
Do you use OpenRouter? Well, if you do, NEAR AI is now on there. You can start today with private inference through OpenRouter uh really, really easily. And I'm going to show you how we do that. I already logged into my account and I'm going to go over to models. And you don't search here by provider.
- 0:17
I'm going to search for the actual models. Right now, uh GLM uh 5.3 Flash is available through NEAR AI. Again, this is leveraging private inference. So, uh all of your queries are protected. And if I click into here, you can see right here NEAR AI is listed, uh, I can set it. So I use this endpoint directly if I'm using OpenRouter in any of my products, which is really cool. So I can just click on that and it will give me the code uh, snippet. It will tell me to only use NEAR AI. Uh, but
- 0:50
you probably should have some fallbacks just in case. That's always good to have redundancies. And you can see here uh what the throughput is, what the uptime is since it got started. Uh it's only been here for a couple days and you can get going and it's great and you can use all this stuff. And I think the really important piece right now with privacy, everything is right here. Uh prompt training, no
- 1:13
data retention, zero retention. Uh moderation obviously is still up to the developer who's leveraging this obviously. Uh but the provider here NEAR AI cannot see any of your queries which is really great and so all your stuff is protected and it's private through TEEs.
- 1:30
So if you haven't already and you're using OpenRouter check it out. So if you have not started using OpenRouter yet for different models, try it out today and you can use NEAR AI directly.