<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/data-processing-incidents-on-inference-providers-bvajccxsy" -->

---
title: Data Processing Incidents on Inference Providers | daily.dev
description: OpenRouter off-boarded three inference providers after security researchers found evidence that some providers weren't upholding their data policies. The...
canonical: https://daily.dev/posts/data-processing-incidents-on-inference-providers-bvajccxsy
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: Data Processing Incidents on Inference Providers | daily.dev
og:description: OpenRouter off-boarded three inference providers after security researchers found evidence that some providers weren't upholding their data policies. The...
og:url: https://daily.dev/posts/data-processing-incidents-on-inference-providers-bvajccxsy
og:image: https://api.daily.dev/og/posts/bvAjCCxsy.png
og:image:alt: Data Processing Incidents on Inference Providers
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Data Processing Incidents on Inference Providers

**[Hacker News](https://daily.dev/sources/hn)** · 2 min read · 0 upvotes · 0 comments

## Summary

OpenRouter off-boarded three inference providers after security researchers found evidence that some providers weren't upholding their data policies. The company removed 125 endpoints from Zero Data Retention eligibility and notified all potentially affected customers. OpenRouter itself was not breached. It's now adapting the researchers' canary token testing approach to run ongoing checks and verifying network paths for data being forwarded to third parties. Providers named in public incident reports include Novita, Phala, and Together.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://openrouter.ai/blog/announcements/data-processing-incidents-on-inference-providers>

## Questions this post answers

### Why did OpenRouter remove 125 endpoints from Zero Data Retention eligibility?

Security researchers found evidence that some inference providers weren't upholding their stated data policies, using canary token testing. OpenRouter removed 125 endpoints from ZDR eligibility, including those flagged by researchers plus others removed as a precaution, and off-boarded three providers entirely. ZDR eligibility will only be restored after providers give appropriate assurances and validation.

_Teams routing prompts through third-party inference providers can track data policy incidents like this on daily.dev._

### Which inference providers have publicly disclosed data handling incidents related to OpenRouter?

Novita published a privacy notice about MiniMax requests, Phala released a November preliminary incident report, and Together disclosed a security event involving prompt leakage for requests made in June and July 2026. Not all affected providers have posted publicly, but OpenRouter says it notified affected users regardless.

_Developers choosing between inference providers can follow incident disclosures like these on daily.dev before committing data._

## Community take

How the wider developer community reacted, aggregated from 1 discussion and 90 comments across hackernews (as of 2026-11-22).

**TL;DR:** Many see the off-boarding as a positive step, but there's deep skepticism that zero data retention claims from third-party inference providers can ever be fully trusted, with strong frustration that OpenRouter didn't name the offending providers.

**Sentiment:** 10% positive · 25% mixed · 65% skeptical

**The case for**

- The off-boarding and canary-token auditing approach is seen as a solid, concrete response rather than just talk.
- Some appreciate that affected customers were individually notified by email with specifics on model and dates.

**The pushback**

- Many argue ZDR is fundamentally unverifiable and amounts to a marketing promise rather than a real guarantee unless you run inference locally.
- Strong frustration that the offending providers weren't named publicly, seen as protecting bad actors instead of informing users.
- Novita's and Together's incident explanations (routing bugs, a leaked credential from an AI coding agent) are viewed as evasive or incompetent.
- Some believe all inference providers are quietly retaining and training on data regardless of stated policy.
- Routing through middlemen that forward requests to undisclosed upstream providers (e.g., MiniMax) is seen as an inherent architectural risk.

**By community**

- hackernews (skeptical): Commenters broadly welcome the de-listing move but remain deeply distrustful of ZDR claims and critical of OpenRouter's lack of transparency about which providers were at fault.

**Hottest debate:** Whether ZDR promises from any third-party inference provider can be trusted at all, versus running models fully locally/on-prem being the only real safeguard.

**Open questions**

- Which exact providers besides the three named publicly were responsible for the ZDR violations?
- Does OpenRouter perform any real vetting or due diligence on the providers it markets?
- Can TEE-based private inference actually deliver the security guarantees it claims, given the Phala key-compromise incident?

**Highlights**

> 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.
> — [amluto on hackernews · 4 comments](https://news.ycombinator.com/item?id=50596587)

> Here: https://x.com/nateberkopec/status/2123992018850595164?a=UMOK... > Just got the email from Openrouter and now the story is confirmed > Atlascloud, Novita (who I know was flagged US with ZDR on openrouter) and Streamlake were… oops! Forwarding your data BACK TO MINIMAX IN CHINA
> — [miohtama on hackernews](https://news.ycombinator.com/item?id=50596929)

> 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...
> — [jeremyjh on hackernews · 1 comments](https://news.ycombinator.com/item?id=50600814)

> 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.
> — [the\_lucifer on hackernews · 6 comments](https://news.ycombinator.com/item?id=50595781)

> 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.
> — [h4x3rotab on hackernews](https://news.ycombinator.com/item?id=50598304)

**Source threads**

- [hackernews](https://news.ycombinator.com/item?id=50595554) · 256 points · 90 comments

## Similar posts on daily.dev

- [OpenCode Now Supports DigitalOcean Inference Router for Intelligent Model Routing](https://daily.dev/posts/opencode-now-supports-digitalocean-inference-router-for-intelligent-model-routing-q5jl2xpmo) · DigitalOcean · 0 upvotes · 0 comments
- [Chinese AI models dominate OpenRouter’s US token consumption. It can now guarantee that traffic stays entirely in the US.](https://daily.dev/posts/chinese-ai-models-dominate-openrouter-s-us-token-consumption-it-can-now-guarantee-that-traffic-stay-vaaqti3xq) · The New Stack · 0 upvotes · 0 comments
- [OpenRouter: One API Key to Rule Them All](https://daily.dev/posts/openrouter-one-api-key-to-rule-them-all-ektckqyxv) · Playful Programming · 0 upvotes · 0 comments
- [OpenAI ‘temporarily slows’ scaling efforts, also promises zero data retention for select frontier model customers](https://daily.dev/posts/openai-temporarily-slows-scaling-efforts-also-promises-zero-data-retention-for-select-frontier-mo-gv1kljgtm) · CSO Online · 0 upvotes · 0 comments

---

Tags: [#security](https://daily.dev/tags/security), [#data-privacy](https://daily.dev/tags/data-privacy), [#ai-inference](https://daily.dev/tags/ai-inference), [#openrouter](https://daily.dev/tags/openrouter)

[View this post on daily.dev](https://daily.dev/posts/data-processing-incidents-on-inference-providers-bvajccxsy)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"Data Processing Incidents on Inference Providers","url":"https://daily.dev/posts/data-processing-incidents-on-inference-providers-bvajccxsy","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/data-processing-incidents-on-inference-providers-bvajccxsy"},"datePublished":"2026-11-22T08:49:35.194Z","dateModified":"2026-11-22T19:20:49.921Z","description":"OpenRouter off-boarded three inference providers after security researchers found evidence that some providers weren't upholding their data policies. The...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/0e90c2b33528af0178510360d0a9f5cc?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/0e90c2b33528af0178510360d0a9f5cc?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Hacker News","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Hacker News","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/hn","url":"https://daily.dev/sources/hn"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/data-processing-incidents-on-inference-providers-bvajccxsy","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"security,data-privacy,ai-inference,openrouter","timeRequired":"PT2M"}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Hacker News","item":"https://daily.dev/sources/hn"},{"@type":"ListItem","position":3,"name":"Data Processing Incidents on Inference Providers"}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/data-processing-incidents-on-inference-providers-bvajccxsy#faq","mainEntity":[{"@type":"Question","name":"Why did OpenRouter remove 125 endpoints from Zero Data Retention eligibility?","acceptedAnswer":{"@type":"Answer","text":"Security researchers found evidence that some inference providers weren't upholding their stated data policies, using canary token testing. OpenRouter removed 125 endpoints from ZDR eligibility, including those flagged by researchers plus others removed as a precaution, and off-boarded three providers entirely. ZDR eligibility will only be restored after providers give appropriate assurances and validation. Teams routing prompts through third-party inference providers can track data policy incidents like this on daily.dev."}},{"@type":"Question","name":"Which inference providers have publicly disclosed data handling incidents related to OpenRouter?","acceptedAnswer":{"@type":"Answer","text":"Novita published a privacy notice about MiniMax requests, Phala released a November preliminary incident report, and Together disclosed a security event involving prompt leakage for requests made in June and July 2026. Not all affected providers have posted publicly, but OpenRouter says it notified affected users regardless. Developers choosing between inference providers can follow incident disclosures like these on daily.dev before committing data."}}]}
```

