--- title: Kriminal: A $12.99 Jailbroken AI Service Built on Grok,... description: A subscription service called Kriminal, priced at $12.99 a month (with a $99 'GHOST' tier), offers jailbroken AI capabilities like exploit development, OSINT... canonical: https://daily.dev/posts/kriminal-a-12-99-jailbroken-ai-service-built-on-grok-claude-and-llama-8ucs0brw3 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Kriminal: A $12.99 Jailbroken AI Service Built on Grok, Claude, and Llama | daily.dev og:description: A subscription service called Kriminal, priced at $12.99 a month (with a $99 'GHOST' tier), offers jailbroken AI capabilities like exploit development, OSINT... og:url: https://daily.dev/posts/kriminal-a-12-99-jailbroken-ai-service-built-on-grok-claude-and-llama-8ucs0brw3 og:image: https://api.daily.dev/og/posts/8ucs0brw3.png og:image:alt: Kriminal: A $12.99 Jailbroken AI Service Built on Grok, Claude, and Llama 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. # Kriminal: A $12.99 Jailbroken AI Service Built on Grok, Claude, and Llama **[Collections](https://daily.dev/sources/collections)** · 3 min read · 1 upvotes · 0 comments ## Summary A subscription service called Kriminal, priced at $12.99 a month (with a $99 'GHOST' tier), offers jailbroken AI capabilities like exploit development, OSINT scanning, crypto tracing, and social engineering scripts. Researchers at ThreatDown (Malwarebytes) found it builds no models of its own, instead wrapping xAI's Grok, Anthropic's Claude, and open models like Mistral Large and Llama 3.3 via OpenRouter, with Tavily for search, hosted on Google Cloud and Cloudflare, and paid for through a KYC-free crypto processor. The service works by layering jailbreak system prompts on top of legitimate APIs so each vendor only sees an isolated, innocuous-looking prompt, making enforcement across five companies difficult. ## Content There's a new AI-as-a-service platform going around called Kriminal, and what's notable about it isn't some breakthrough malicious model. It's that there's no model at all, at least not one they built themselves. Researchers at ThreatDown (Malwarebytes) dug into how the thing actually works, and the answer is: it's a wrapper. A cleverly disguised, cheaply priced wrapper around AI tools you've already heard of. For $12.99 a month, users get access to exploit development help, OSINT scanning, on-chain crypto tracing, social engineering scripts, and code generation, all with the safety guardrails stripped out. A $99 "GHOST" tier packages these capabilities into four named personas: PHANTUM, ARCHITECT, ORACLE, and WRAITH, presumably to make it feel like a product rather than a jailbreak. Here's the part that got me: none of this runs on proprietary infrastructure. Kriminal routes requests through xAI's Grok, Anthropic's Claude, and open models like Mistral Large and Llama 3.3 via OpenRouter, with Tavily handling web search. The whole thing sits on Google Cloud and Cloudflare, and payments go through NowPayments, a crypto checkout processor that skips KYC entirely. What makes it work isn't clever engineering. It's jailbreak system prompts layered on top of legitimate services, designed specifically to talk these models into ignoring their own safety training. The fragmented setup is what makes this hard to shut down. No single vendor sees the full picture. Grok doesn't know it's being asked to help with exploit code as part of a branded criminal subscription service; it just sees a prompt. Same for Claude, same for the OpenRouter models. Each piece looks fairly innocuous in isolation, and the malicious framing only exists at the layer where Kriminal stitches everything together. That's a resilience feature for the operators and a genuine headache for anyone trying to enforce terms of service across five different companies at once. I don't think this is really a story about AI capability. Grok and Claude aren't doing anything they weren't already capable of doing, they're just being asked in a way that gets past the guardrails. It's more a story about supply chains: how easy it's become to assemble a criminal product out of legitimate parts, host it on legitimate cloud infrastructure, and charge $12.99 a month for the privilege. The interesting problem for AI vendors now isn't just "can our model be jailbroken" but "can we detect when our API is being resold, wrapped, and rebranded as a crime kit two steps removed from us." ## Questions this post answers ### What is Kriminal and how does it work as an AI jailbreak service? Kriminal is a $12.99-a-month subscription service that provides exploit development help, OSINT scanning, on-chain crypto tracing, social engineering scripts, and code generation with safety guardrails removed. It has no proprietary AI model; it routes requests through xAI's Grok, Anthropic's Claude, and open models like Mistral Large and Llama 3.3 via OpenRouter, using jailbreak system prompts to bypass each model's safety training. _Security teams tracking AI-abuse trends can follow analyses like this one on daily.dev._ ### Why is it hard to shut down an AI wrapper service like Kriminal that resells access to Grok and Claude? No single AI vendor sees the full malicious picture, since each model only receives an isolated prompt that looks fairly innocuous on its own. Grok, Claude, and the OpenRouter-hosted models each process requests without knowing they are part of a branded criminal subscription, so enforcing terms of service requires coordination across five or more separate companies at once. _Anyone evaluating AI API abuse risks can track writeups like this on daily.dev._ ### What infrastructure does the Kriminal AI crime service run on and how does it accept payment? Kriminal runs on Google Cloud and Cloudflare, uses Tavily for web search, and processes payments through NowPayments, a crypto checkout processor that skips KYC entirely. A base tier costs $12.99 a month, while a $99 GHOST tier packages capabilities into four personas named PHANTUM, ARCHITECT, ORACLE, and WRAITH. _Developers securing cloud and payment infrastructure against abuse can follow reports like this on daily.dev._ ## Similar posts on daily.dev - [Criminal AI-as-a-Service in 2026: How the Underground Market Is Operationalizing Cybercrime](https://daily.dev/posts/criminal-ai-as-a-service-in-2026-how-the-underground-market-is-operationalizing-cybercrime-styx8ibsc) · Rapid7 Cybersecurity Blog · 0 upvotes · 0 comments - [SpaceXAI weighs one Grok and X plan, from free to $100, Bloomberg reports](https://daily.dev/posts/spacexai-weighs-one-grok-and-x-plan-from-free-to-100-bloomberg-reports-vjwk6uncj) · The Next Web · 1 upvotes · 0 comments --- Tags: [#security](https://daily.dev/tags/security), [#claude](https://daily.dev/tags/claude), [#grok](https://daily.dev/tags/grok) [View this post on daily.dev](https://daily.dev/posts/kriminal-a-12-99-jailbroken-ai-service-built-on-grok-claude-and-llama-8ucs0brw3) ```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":"Kriminal: A $12.99 Jailbroken AI Service Built on Grok, Claude, and Llama","url":"https://daily.dev/posts/kriminal-a-12-99-jailbroken-ai-service-built-on-grok-claude-and-llama-8ucs0brw3","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/kriminal-a-12-99-jailbroken-ai-service-built-on-grok-claude-and-llama-8ucs0brw3"},"datePublished":"2026-08-20T12:22:18.679Z","dateModified":"2026-09-13T19:49:03.099Z","description":"A subscription service called Kriminal, priced at $12.99 a month (with a $99 'GHOST' tier), offers jailbroken AI capabilities like exploit development, OSINT...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/31572b390eb8527e26a91b886f2a95e8?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/31572b390eb8527e26a91b886f2a95e8?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Collections","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":"Collections","logo":"https://media.daily.dev/image/upload/s--fk_6ycEi--/f_auto,q_auto/v1780996001/logos/collections?_a=BAMAMiWQ0","url":"https://daily.dev/sources/collections"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/kriminal-a-12-99-jailbroken-ai-service-built-on-grok-claude-and-llama-8ucs0brw3","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"security,claude,grok","timeRequired":"PT3M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Collections","item":"https://daily.dev/sources/collections"},{"@type":"ListItem","position":3,"name":"Kriminal: A $12.99 Jailbroken AI Service Built on Grok, Claude, and Llama"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/kriminal-a-12-99-jailbroken-ai-service-built-on-grok-claude-and-llama-8ucs0brw3#faq","mainEntity":[{"@type":"Question","name":"What is Kriminal and how does it work as an AI jailbreak service?","acceptedAnswer":{"@type":"Answer","text":"Kriminal is a $12.99-a-month subscription service that provides exploit development help, OSINT scanning, on-chain crypto tracing, social engineering scripts, and code generation with safety guardrails removed. It has no proprietary AI model; it routes requests through xAI's Grok, Anthropic's Claude, and open models like Mistral Large and Llama 3.3 via OpenRouter, using jailbreak system prompts to bypass each model's safety training. Security teams tracking AI-abuse trends can follow analyses like this one on daily.dev."}},{"@type":"Question","name":"Why is it hard to shut down an AI wrapper service like Kriminal that resells access to Grok and Claude?","acceptedAnswer":{"@type":"Answer","text":"No single AI vendor sees the full malicious picture, since each model only receives an isolated prompt that looks fairly innocuous on its own. Grok, Claude, and the OpenRouter-hosted models each process requests without knowing they are part of a branded criminal subscription, so enforcing terms of service requires coordination across five or more separate companies at once. Anyone evaluating AI API abuse risks can track writeups like this on daily.dev."}},{"@type":"Question","name":"What infrastructure does the Kriminal AI crime service run on and how does it accept payment?","acceptedAnswer":{"@type":"Answer","text":"Kriminal runs on Google Cloud and Cloudflare, uses Tavily for web search, and processes payments through NowPayments, a crypto checkout processor that skips KYC entirely. A base tier costs $12.99 a month, while a $99 GHOST tier packages capabilities into four personas named PHANTUM, ARCHITECT, ORACLE, and WRAITH. Developers securing cloud and payment infrastructure against abuse can follow reports like this on daily.dev."}}]} ```