usernomdeguerre 10 hours ago

Wow I missed this story entirely. I'd seriously considered moving to OpenRouter at my next external AI renewal due to the breadth of options and pricing but this has absolutely ended that consideration.

It's clear from the breadth and variety of exposures that there's no due diligence or guardrails inside OpenRouter on who they market and how they classify their providers. Their response to this indicates a few first steps but none seem to have the force of law and i'd have expected more from a Stripe aquisition.

Very glad that my daily driver is a local model and I only use these external services sparingly.

redrix 15 hours ago

Nice to see this move by OR.

That being said, the best ZDR agreement is the one that your company has directly with the inference provider(s). Unfortunately:

- Not everyone is large enough (or has enough $$) for that. - Ultimately you’re in the same position of having to “trust” that they’re doing ZDR; even after auditing their security docs.

  • rkagerer 14 hours ago

    I'd say that's the second-best one. The most robust ZDR policy is running your own AI locally and enforcing the retention policy yourself.

    • az226 14 hours ago

      This right here.

      ZDR is a marketing ploy. It’s not real. Only naive people believe it’s real.

      • fc417fc802 8 hours ago

        I generally trust (for example) bedrock to honor the ZDR agreement as written and to have secure infrastructure. However I don't fully trust the routing service I use to access bedrock to have rigorous security practices.

    • dash-44 5 hours ago

      Yes that's why Revolut (Europe's largest digital bank) has just started self hosting open weight models using 56 Nvidia B200 GPUs. So their 2000 engineers will use that for most things.

      Expect many more companies to follow.

      • JimDabell 3 hours ago

        That’s not true. They switched for cost and productivity reasons.

        They explicitly specified USA, EU, UK, and Israel-hosted services as acceptable as long as it’s ZDR. They said that inference hosted in China was something they didn’t want.

        The talk is here: https://www.youtube.com/live/akEqiXFo3GM

  • msdz 13 hours ago

    > Ultimately you’re in the same position of having to “trust” that they’re doing ZDR

    Could it be a moat to offer a zero-trust (based on technical impossibility) ZDR solution for corporate customers, or is everyone who’s paranoid enough about that just going to host local models on-prem eventually?

    • edoceo 12 hours ago

      Not an expert, I consult with medium size firms that so far have seen nice things from the AI. However, the lightly-informed C-Suite hears about the hacks and the data threats on CNBC/Bloomberg which has made them cautious. Even with their cursory understanding they are saying "can we look at these open models". N=2

      • TeMPOraL 6 hours ago

        Isn't that the problem themselves, in that those issues only exist if you go for open models, but outsource inference to third parties? It seems to me that there are only two safe options: either go full local and do inference on-prem/in your own cloud, or just deal with the big labs directly. Everything between these two extremes involves third parties with sketchy reputations.

    • 392 11 hours ago

      I think there's definitely a market, on prem feels expensive and intimidating for many no matter what it actually costs or takes

      • otabdeveloper4 3 hours ago

        "On prem" here just means renting a server with a GPU. (Possibly "in the cloud".)

        • Geezus_42 3 hours ago

          So even more expensive.

  • the_lucifer 12 hours ago

    > Nice to see this move by OR.

    Unfortunately, I highly disagree. I'm deeply disappointed that they don't even have the gall to call the violators out so we can be informed. Dug around for so long and I can't still find all the rule-breaking providers. It's a bunch of "trust us bro" nonsense. Name and shame should be the minimum here.

    • jeremyjh 1 hour ago

      If they are going to name and shame they have to put in a lot more work first to be sure their evidence is tight enough to dissuade or dismiss defamation lawsuits, and there would still be risk and lots of drama for people too busy for that shit. It just isn't realistic to expect that in my opinion.

      Its possible to be very certain of the evidence you have but not be able to prove its not been tampered with or fabricated, for example.

  • doctorpangloss 11 hours ago

    what move? OpenRouter is garbage, how could Stripe have thought it was a good idea to buy them? $7b for revenues that will evaporate, inevitably, all at once, in weeks!

a012 13 hours ago

I was affected with just one prompt by this incident, which is non issue to me because I was trying different models at that time so there was nothing serious. The troubling problem I see in this incident is the “providers” in OR just casually forwarding prompts to upstream (MiniMax) instead of providing inference service them selves.

Coincidentally later on I set my OR settings to only allow some providers that I can rely on, not blindly let OR route to wherever they like.

  • yard2010 8 hours ago

    For those who don't know this is the first thing you have to do when using openrouter, change guardrails settings from blacklist to whitelist, pick 1-2 providers you trust (do your homework, or let $agent do it for you), and if you want to go the extra mile pick the fastest/cheapest provider for you model (again do the research). The defaults can't be trusted if you care about privacy.

    • disgruntledphd2 5 hours ago

      The first thing I did was turn on ZDR, and then looked for region hosted models. Unfortunately this is gated behind a business plan.

      • danvdb 18 minutes ago

        You can use something like Requesty (requesty.ai) where whitelisting is the default and in-region inference is available in every plan.

codedokode 2 hours ago

This reminds again that all "promises" of companies are a lie and you better not trust their "privacy policies".

  • jeremyjh 1 hour ago

    They are not all lies, but these are definitely circumstances in which case lies should be expected because it is possible to profit significantly by making them. Lots of new companies springing up over night making promises they won't steal the valuable data passing through their hands.

    The number of providers they have added is absurd, there have to be some that have almost no compute and just upstream everything. So stealing is their only possible business model.

    openrouter.ai hasn't really promised that they do any vetting or due diligence, as far as I know. This whole space has had the scammy crypto vibe from the beginning.

smuhakg 13 hours ago

I got the email. The tl;dr is MiniMax (the Chinese AI lab) is actively reading your prompts, and a security researcher discovered this by putting decoy URLs into their requests which were accessed by MiniMax.

More importantly, many providers (AtlasCloud, Novita and StreamLake were named but there seems to be more implied in the blog post) have been rerouting requests to MiniMax servers on the backend and not processing their own prompts.

The obvious question is whether OpenRouter has done any auditing on their inference providers. Based on the researcher's blog post, I guess not.

https://www.dragos.com/blog/ai-supply-chain-security-canary-...

  • amluto 11 hours ago

    Novita’s blog post is unreal:

    https://novita.ai/blog/minimax-requests-privacy-notice/

    > A routing configuration issue caused an API forwarding path intended for specific customers to apply more broadly.

    > If you used MiniMax models through Novita between 2026-05-28 and 2026-11-11, your request contents may have been accessed or retained by a third party. This includes requests made directly through Novita and those routed to Novita through OpenRouter.

    “routing configuration issue”?!? Imagine if AWS got caught storing your S3 on some other provider’s disks and they said it was a “routing configuration error”! They shouldn’t have software that can do this at all! The API is supposed to be for their own GPUs!

    Together’s is weird:

    > As part of routine software testing, we collected prompt samples authorized for that use and stored them in a private Hugging Face repository. The samples came from requests to the MiniMax M3, GLM-5.2, and Kimi K2.6 models, primarily on June 24, July 4, July 8, and July 14, 2026.

    Did Dragos really “authorize” Together to store their canary “prompt samples”? Somebody screwed up here.

    • speedping 11 hours ago

      You left out the best part from Together’s “During development of our public GitHub repository, xorl, an AI coding agent committed a credential that provided access to the private Hugging Face repository”

      • g-b-r 9 hours ago

        So surprising that these services are vibecoded

        • gessha 2 hours ago

          “I’m shocked! Shocked that gambling is happening here!”

          The only thing you can trust is possibly(and not always) your own hardware, your own staff and companies you can directly inspect.

    • luke-stanley 2 hours ago

      This and the way Together seems to be framing it in their disclosure post ( https://www.together.ai/blog/security-event-prompt-leakage-f... ) raises a lot of serious questions about the integrity of Together in more ways than one. Is the Together post really a fair disclosure that accounts for what Dragos identified? Their explanation is far from clear. It makes me wonder what the legal agreement was, and if the work mentioned in Together's "Polymorphic Inference" blog post, mentioning speculative decoding (using cheap models to draft) is a related practice they have routinely. Pretty worrying. I'm sure some will make this into simply blaming Chinese companies for this (not you), which wouldn't fairly account for the risks of the industries optimisation incentives. (Together's HQ is in SF AFAIK).

    • jruohonen 2 hours ago

      > Somebody screwed up here.

      Yet, the basic idea of these myriads of canaries is good. I see several use cases, including in research.

    • jeremyjh 1 hour ago

      Together.ai has made a public announcement that they are fucking clowns who cannot be trusted with anyone's data.

      > During development of our public GitHub repository, xorl, an AI coding agent committed a credential that provided access to the private Hugging Face repository. Our automated scanning detected the exposed credential, and we reverted the commit.

      They didn't rotate the credential they knew was exposed!?! I guess "an AI coding agent" didn't think of doing that. Also FYI kids, reverting a commit does not delete the data from the repository.

      In before they claim an AI agent wrote this disclosure that included certain "errors".

      https://www.together.ai/blog/security-event-prompt-leakage-f...

      • amluto 14 minutes ago

        This is real classy coming from a neocloud with a $8.3B valuation. Shouldn’t a neocloud with a bunch of Nvidia GPUs and a bunch of investment (including from Nvidia) be using those Nvidia GPUs to run inference in house? Isn’t that the entire point?

        Amusingly, the first time I typed “neocloud” it got autocorrected to “bro cloud”. I almost decided to leave it alone.

  • odie5533 9 hours ago

    Is Minimax basically an untrusted scammer at this point?

    • luke-stanley 2 hours ago

      This explanation doesn't account for TogetherAI's sharing canary tokens!

leobg 10 hours ago

As an aside, I always wondered about decision models. They are so cheap. I can’t help but think that providers must be using them as honeypots for data acquisition:

You do dont find real world company documents on the web that you could train on. And companies certainly won’t upload them to you so you can improve your models, not even with incentives. But provide a model with dirt cheap input pricing and free outputs - and you gut yourself a firehose.

To think that the “labs” wouldn’t train on this is like thinking they’d never have trained on shadow library books because, “well, it’s illegal”. They did it back then and it gave them an advantage over the competition. So why wouldn’t they now?

  • jofzar 9 hours ago

    > As an aside, I always wondered about decision models. They are so cheap. I can’t help but think that providers must be using them as honeypots for data acquisition:

    This one is actually the simplest for me, you can run decision models on your own laptop, and you can see how cheap and fast they are, it's actually amazing how "good" (jev etc is still better) they are.

    So instead of thinking of them as a "LLM" for hosting pricing, it's easier to think of them as a API as a service, it's practically "how much can we charge someone for processing in bulk for an action and to get them sticky".

    You setup whatever workflow you want for the decision models and then you practically never change it, because why would you, it's dirt cheap (for you and they provider), they (the provider) is making money on your outsourcing for every API action and it's going to be consistent because it's going to be for a long time.

    • pyinstallwoes 8 hours ago

      But nobody knows how jev is actually built , what open source decision models? All the ones available kind of jerryrig it on top of a LLM last I looked into it.

      • ytxm 7 hours ago

        IIRC JEV is known to be built on Qwen, you simply ask it questions to that effect. No idea where I saw the thread, but someone did this exhaustively using many variations and the answer was conclusive

  • janalsncm 9 hours ago

    Decision models can be cheap because it’s a fundamentally cheaper technology than LLMs. So that part isn’t suspicious at all to me.

    That said, if you don’t believe OpenAI when they say they don’t train on your decision API inputs you shouldn’t trust them regarding LLM inputs either.

mbgerring 13 hours ago

You should assume that no inference provider is honoring any public commitment they have made about data retention. You should assume that 100% of what you share with an LLM over the wire is being recorded, saved, and used as training data. Any other stance is indefensible naïveté for anyone who has been around the tech industry for more than, let’s say, 5 years.

  • LeBit 13 hours ago

    That’s why I never use a real email or real name.

    The payment is where you have to provide real infos.

    I just hope the payment infos is stored at a different place than prompt/conversations storage so mapping of user id to payment id is not trivial.

    • octoberfranklin 12 hours ago

      The payment is where you have to provide real infos.

      What? I've never provided real world info when paying.

      I always use Monero when paying for AI tokens. Crazy to do anything else.

      • eddythompson80 11 hours ago

        More insane to use cloud provider AI. Do you care about security or privacy at all? I run my own models. Crazy to do anything else.

        • kqr 10 hours ago

          How do you prevent the model from exfiltrating data? Are you running it in a harness with no internet access or doing something more sophisticated?

          • exe34 9 hours ago

            Don't give it internet access.

      • actualeff0rt 6 hours ago

        Which inference providers lets you pay with Monero?

  • Fordec 12 hours ago

    I fully expect them to be storing it, just the period of time they are storing it for is up for debate. Even everything older than six months ago, at this stage for me would be pure junk. If you want to train on that when I was still using soul files and barely any harness, heh, well that's your problem now!

  • kqr 11 hours ago

    There are real exceptions to this rule, like Tinfoil.sh who cryptographically ensure they cannot retain the data. (As long as one trusts the chipmaker, I suppose.)

    Of course, even with a privacy-respecting inference provider, if a modern, relentlessly proactive LLM runs in a harness with internet access, it can easily accidentally exfiltrate your data anyway.

    • g-b-r 9 hours ago

      It seems great

    • fc417fc802 8 hours ago

      > As long as one trusts the chipmaker

      Rather you have to trust the equipment operator (which in this era of clouds isn't necessarily the vendor) not to intentionally compromise their own hardware (see https://tee.fail/). But other than that quibble yeah, tinfoil looks like an excellent option that's much less likely to leak than current alternatives.

  • dhakiptk 10 hours ago

    I think tinfoil provides that(if CPU manufacturer can be trusted)

  • codeulike 9 hours ago

    Mistral are bound by EU law

    • ozgrakkurt 9 hours ago

      All citizens are bound by law but this doesn't mean you will have 100% confidence while walking in a dark alley at 3AM in a city you don't know.

    • sheo 7 hours ago

      North Koreans that could hack Mistral and sell data for crypto are not.

  • metalspot 4 hours ago

    read anthropics tos. it says very specifically that they do not train on "inputs" and "outputs" of customer data. everything in-between is openly declared to be their legal right to train on. in terms of their ability to extract intelligence from inference there is zero difference between this and training on "inputs" and "outputs."

    • ActionHank 2 hours ago

      By the definition of how these companies see data, a simple transform to restructure the input and output would be sufficiently different to pass this bar.

      • metalspot 2 hours ago

        exactly. the tokenization is not the input.

hasyimibhar 9 hours ago

ZDR through AWS Bedrock is still safe right? AFAIK Bedrock self hosts all models they provide. Even proprietary weight models like Claude are run within AWS network.

  • cedws 8 hours ago

    Yes, however some models don’t allow ZDR IIRC. I think Fable requires enabling data retention.

  • happosai 7 hours ago

    Assuming you trust AWS is not, say tapped by NSA. Of course, if your data is on AWS already, this makes no difference.

  • 0xbadcafebee 3 hours ago

    Bedrock provides an incredibly limited and old set of models. Anthropic and OpenAI models are not hosted by AWS they are passed through.

the_lucifer 14 hours ago

Is there any information on the three companies that violated ZDR? I could only find mention of one of those affected providers in the original article (that OR links in their post): Lark. Trying to figure out if I ended up routing to one of the affected "ZDR" endpoints.

I don't understand why these names are being coyly tiptoed around as if they're trying to protect the privacy of these companies blatantly lying about ZDR. Why isn't OR calling out the actual providers they off-boarded by name? Why is OR basically protecting them under a cover of anonymity?

Lost a lot of trust in OR that's for sure, in-spite of some other positive reception I see here for this move.

Edit: Wording update.

  • fedpost 13 hours ago

    Given how trust based the LLM world has to be right now, there should be zero tolerance for anyone doing stupid shit like this

  • SwellJoe 13 hours ago

    They state they notified customers who used any of the problem models via email (I got an email about it), which answers your question of whether you ended up routing to one of those endpoints. They apparently know who was routed to all those providers, so you don't have to guess.

    That still leaves the desire of people to make informed decisions about what providers to use unresolved. If reputation is the only thing we have to go on when choosing a provider to send a bunch of sensitive data to, it'd be nice if the folks who actually know whether a provider is trustworthy would share that info.

    But, since the email they sent identified the provider and the model (MiniMax, in my case) and the dates of use, it seems like it won't be secret for long what the other problem providers are. I'll assume the email to users of other providers will also include those details, so people can post about it even if OpenRouter is cagey about it in their blog post.

    • the_lucifer 12 hours ago

      > OpenRouter is cagey about it in their blog post.

      That's exactly what's confusing me though? Why is OR basically covering for these providers by not naming them?

      • jofzar 9 hours ago

        So they don't get sued, you bet this was reviewed by lawyers.

        • ActionHank 2 hours ago

          Also, naming names makes it real and bigger than this reads now with its vagaries.

  • octoberfranklin 12 hours ago

    Is there any information on the three companies that violated ZDR?

    I have my harness set to show me "diffs" of the catalog when I start a new session.

    There was a mass-downgrade to non-ZDR a few days ago. At the time I wondered why.

    It was all older (not latest release) Chinese models being run by second-party inference providers.

angry_octet 6 hours ago

Model routing was always insane. Until there is attestation from the providers and encryption of the session the temptation to peek will be too great.

Ungh 9 hours ago

As usual, if you really give a shit about your data, don’t give it to someone else.

solarkraft 14 hours ago

With this change, everything is good in the world and we can just take everyone’s word that they’re actually ZDR without any audits or technical enforcement again. Right?

cobychapple 13 hours ago

Zero data retention, give or take 125 endpoints.

neuroelectron 4 hours ago

Of course, the entire point of running an inference provider is stealing sensitive proprietary data. What are they gonna do? sue for copyright infringement?

g-b-r 13 hours ago

This revealed something as delightful as a purported TEE provider having their "software TEE" root keys hacked, through "a legacy internet-facing service with management interfaces that relied on loopback access for trust and did not independently authenticate callers" (https://phala.com/posts/phala-november-preliminary-incident-...)

  • fc417fc802 8 hours ago

    > having their "software TEE" root keys hacked

    Did they? It says the attacker was able to create unauthorized CVMs (makes sense), able to obtain application credentials (significance unclear to me), and that they rotated their KMS root key in response (which is confusing to me because if they use hardware backed keys then I don't understand how the root key could ever be compromised).

    So does this imply they aren't actually providing a proper hardware backed TEE? But that's also confusing because at least if you provision a VM you can have the hardware attest to its state.

    • g-b-r 6 hours ago

      Yeah, for what I can tell (I only wasted a little time in looking into it), they say TEE TEE TEE but then oops but TEE could be compromised, it locks your data, it's inconvenient, no let's add our better software layer (where those "KMS" root keys are stored).

      (https://docs.phala.com/dstack/design-documents/decentralized... , https://docs.phala.com/dstack/design-documents/key-managemen...)

      • fc417fc802 6 hours ago

        Ah I see, a classic case of a secure system only being as robust as the software you intentionally run on it. New crypto fails in exciting ways after all. At least it's FOSS. https://github.com/Dstack-TEE/dstack

        Actually it looks like a pretty decent solution once the inevitable bugs and footguns are ironed out.

        • h4x3rotab 6 hours ago

          Thanks for the kind words. Any single point of failure must be handled with great care. In this case, the blockchain-based authentication is proven to be more robust than the central one. And there are a lot more to harden. It’s a lesson we learned in a hard way.

    • h4x3rotab 6 hours ago

      Cofounder of Phala here. Happy to explain it.

      The TEEs are real. The problem was caused by software bugs. Running a TEE app (CVM) is more than dropping a binary into TEE. Two problems remain: (1) migration, you don’t want to completely lose the data if the hardware dies, and (2) upgradability, you want to patch the app when there’s a bug without building the state from scratch. So we want to decouple the states from hardware.

      Using the hardware key is painful. The solution is to introduce a separate key manager (KMS). It releases keys to the app iff. the app follows a predefined “policy” (code, hardware attestation, etc). We can decouple the states from hardware based on a remote KMS.

      This is rather a common design. You can find it in our system, as well as other mature frameworks like Confidential Containers.

      As the system grows, a bug was introduced to the policy of how we replicate a KMS’s key material (for redundancy). The attacker exploited it with multiple sophisticated attacks, eventually being able to cheat a running KMS to hand over the keys to a fake one, like it’s a legit KMS replica. As a result, the keys protected by the KMS became at risk.

      To mitigate the problem, the legacy KMS was dropped and a new KMS was deployed. We have then rotated all the keys associated with the legacy KMS. A series of protocol and code audit is ongoing to prevent future failure.

Havoc 14 hours ago

Solid move.

Though frankly I would think anyone even slightly bothered about ZDR does not start their train of thought with "OK so we're going to put a middle man in"

  • Hizonner 1 hour ago

    If you're not an "enterprise", your only way to get even purported ZDR with anything close to frontier models is to use a router. The labs will not deal with individuals, and I suspect in practice it's also too hard for small businesses.