Scammers Profile Link:https://cracked.st/MrPoisonLab
Sales Thread:https://cracked.st/Thread-%E2%9C%85-CLAUDE-CODE-API-SERVICE-%E2%AD%90-ORIGINAL-MODELS-%E2%9A%A1-93-CHEAPER-THAN-ANTHROPIC--2111238
Amount or Item(s) Scammed: 30.32$
Screenshots of Communication:https://imgur.com/a/Wl7DpoG
Additional Information:I purchased a package for a Claude API service advertised at $29. In reality, I was charged approximately $30.32. Everything seemed fine at first. I connected the API to Claude and was told that the package included 500 requests per day.
I started using the service and submitted a single prompt. About 10 minutes later, I received a notification stating that I had already reached my daily limit. During that session, Claude generated several completely random files that I had never requested.
When I contacted the seller, he claimed that I had reached the limit. However, he also provided a website where customers could track their request usage. According to that website, I had only used around 200 requests in total. When I questioned this discrepancy, he told me that the website does not accurately reflect usage because requests made with Opus are counted differently and may count as multiple requests. In other words, the website showed around 200 requests, while the application claimed I had already reached the 500-request limit.
I also noticed that approximately 5.4 million tokens had supposedly been consumed for a very short interaction that produced essentially no useful results. It is extremely difficult to believe that such a simple conversation could consume that many tokens while simultaneously reaching the full daily limit. Throughout the entire process, Claude frequently displayed "Try Again" and API error messages.
I reviewed the sales page and the discussion thread created by the seller. Nowhere did I find a clear statement that purchases were non-refundable. After seeing the excessive token usage and the poor performance of the service, I requested a partial refund. I asked that the cost of the tokens I had already used be deducted and that the remaining amount, approximately $26, be refunded.
The seller refused and claimed that the issue was my lack of knowledge about how to use the Claude API. I later discovered that another customer had reported a very similar problem with the same service.
I am requesting that my money be refunded as soon as possible. I no longer wish to use this service because it has been extremely slow, unreliable, and full of issues. The seller now claims that refunds are not offered, yet this was never clearly disclosed before the purchase. I do not believe this is fair or transparent business practice.
Crypto address for a refund:
0xEbEeb951d0161ddB339d248548d9543297e6fe79
I explained to customer that the count on the website doesn't show the 2.5x multiplier of opus 4.8 I have multiple good reviews on telegram of people that are more than happy with the services. This guy has no idea how API Tokens and say I'm scamming. I'm not refunding and happy to send any mods the chat export.
This post is by a banned member (VelvetV7) - Unhide
19 June, 2026 - 07:46 PM(This post was last modified: 19 June, 2026 - 07:53 PM by VelvetV7. Edited 1 time in total.)
Reply
I had a very poor experience with this seller and feel like the pricing/token system is highly misleading.
The seller claims the service is “93% cheaper than Claude,” but in practice the token usage appears to be inflated by roughly 350×–700×. For example, a simple “Test” prompt consumed around 70,000 tokens, which feels completely unreasonable for such a small input.
There also seems to be an issue with what looks like double counting, after sending a prompt and having tokens deducted, additional tokens were taken again minutes later without any new user action.
When I told this to the seller, the only response I received was essentially that “this is how it works and everyone does it like this,” but none of this is clearly explained upfront. A breakdown of how tokens are calculated and what users are actually paying for would have been important context, especially since most users would reasonably assume the pricing refers to standard Claude-equivalent tokens.
Honestly, the way token usage and pricing are explained (or not explained) is really unclear. It’s hard to understand what you’re actually being charged for, and it doesn’t really match what the marketing makes it sound like.
The website doesn’t give enough detail on how usage is counted, so the billing ends up feeling kind of misleading compared to what you’d expect.
On top of that, the whole setup feels a bit off, especially since refund requests don’t seem to be handled fairly or are basically discouraged, even when the service hasn’t really been used much.
I paid $4.32 for my tokens, and if possible would also love a refund, not sure if I need to make a thread on my own, I only used like 5% of my tokens, so I got no problem with only receiving 95% back, that shouldn't be a problem, right? UPDATE:
The seller just posted in his telegram channel offering "juicy discount" to anyone that has a cracked account. Seems like he will ask them to vouch for a discount, not sure if that's allowed, but definitely shouldn't be.
"if you have a cracked account and bought with us before contact me for a juicy discount on next purchases"
19 June, 2026 - 08:05 PM(This post was last modified: 19 June, 2026 - 10:01 PM by MrPoisonLab.)
Reply
(19 June, 2026 - 07:46 PM)VelvetV7 Wrote: Show More
I had a very poor experience with this seller and feel like the pricing/token system is highly misleading.
The seller claims the service is “93% cheaper than Claude,” but in practice the token usage appears to be inflated by roughly 350×–700×. For example, a simple “Test” prompt consumed around 70,000 tokens, which feels completely unreasonable for such a small input.
There also seems to be an issue with what looks like double counting, after sending a prompt and having tokens deducted, additional tokens were taken again minutes later without any new user action.
When I told this to the seller, the only response I received was essentially that “this is how it works and everyone does it like this,” but none of this is clearly explained upfront. A breakdown of how tokens are calculated and what users are actually paying for would have been important context, especially since most users would reasonably assume the pricing refers to standard Claude-equivalent tokens.
Honestly, the way token usage and pricing are explained (or not explained) is really unclear. It’s hard to understand what you’re actually being charged for, and it doesn’t really match what the marketing makes it sound like.
The website doesn’t give enough detail on how usage is counted, so the billing ends up feeling kind of misleading compared to what you’d expect.
On top of that, the whole setup feels a bit off, especially since refund requests don’t seem to be handled fairly or are basically discouraged, even when the service hasn’t really been used much.
I paid $4.32 for my tokens, and if possible would also love a refund, not sure if I need to make a thread on my own, I only used like 5% of my tokens, so I got no problem with only receiving 95% back, that shouldn't be a problem, right? UPDATE:
The seller just posted in his telegram channel offering "juicy discount" to anyone that has a cracked account. Seems like he will ask them to vouch for a discount, not sure if that's allowed, but definitely shouldn't be.
"if you have a cracked account and bought with us before contact me for a juicy discount on next purchases"
I'm just helping whoever likes my service and want to comment with a vouch, do you really think people would lie over 10% discount if the service was bad? Make sense at least for once
I'm only now having time to really address these 2 guys who clearly have 0 experience with third-party Claude API service
@VelvetV7 has no clue and has never used the official Anthropic API saying my service isn't cheaper, if any of the mods here have used they will be able to tell that the 4$ i charged him for 5M tokens that do last at least 30 minutes of work with Opus 4.8 on 1M context would last 10 minutes or less. He has no right to harm my name by saying I'm misleading anything.
- Also, he doesn't understand how our tokens are calculated which are very different than Anthropic as any person that has used third-party API services know: Input and Output for a "Hi" prompt is the same for a prompt to build an entire website. We rely on Prompt Caching: For example, if you feed the AI a 100,000-token document, the we cache it. On subsequent questions, our model references the cache rather than re-reading the whole text, which reduces the billing cost to Anthropic by up to 90% but still 4$ which is basically our TRIAL will get you done very fast and this guy expected something unrealistic.
@OutCome is another one who has 0 clue of how this works. He bought our Mini Plan which is a monthly subscription with only 500/requests a day for 29$, which gets reset every 24 hours, but is the cheapest package.
He says 500 requests was gone fast but I see his requests for more than 30 minutes of work, he expects the simplest of all packages which I told him that was only suitable if you not heavy coding.
He expected to work many hours with Opus 4.8 on a 1M context and his context window was probably already extensive when he started because it jumped from 0 to 15 requests just to start and then stabilized. Its not a fresh session which will consume more and more requests because the AI when started on an extensive context it has to read history, all docs related to that session, system instructions , prompts and start writing code. Just these will consume a lot of requests.
I never ever made any promisses to these 2 guys and they expected something that was only in their heads.
My website footer and Docs has all the information they said I didn't provide. I don't have to tell every buyer every single rule if they don't ask about it!!
This post is by a banned member (MrPoisonLab) - Unhide
19 June, 2026 - 10:01 PM(This post was last modified: 19 June, 2026 - 10:21 PM by MrPoisonLab. Edited 4 times in total.)
Reply
I'm only now having time to really address these 2 guys who clearly have 0 experience with third-party Claude API service
@VelvetV7 HE HASNT EVEN TRIED USING THE SERVICE, HE USED ONE PROMPT AND CAME TO COMPLAIN, HIS KEY IS STILL FULL OF TOKENS!!
Has no clue and has never used the official Anthropic API saying my service isn't cheaper, if any of the mods here have used they will be able to tell that the 4$ i charged him for 5M tokens that did last for him at least 30 minutes of work with Opus 4.8 on 1M context would last 10 minutes or less with the OFFICIAL API. He has no right to harm my name by saying I'm misleading anything.
- Also, he doesn't understand how our tokens are calculated which are very different than Anthropic as any person that has used third-party API services know: Input and Output for a "Hi" prompt is the same for a prompt to build an entire website. We rely on Prompt Caching: For example, if you feed the AI a 100,000-token document, the we cache it. On subsequent questions, our model references the cache rather than re-reading the whole text, which reduces the billing cost compared to Anthropic by up to 90% but still 4$ which is basically our TRIAL will get you done very fast and this guy expected something unrealistic.
@OutCome is another one who has 0 clue of how this works. He bought our Mini Plan which is a monthly subscription with only 500/requests a day for 29$, which gets reset every 24 hours, but is the cheapest package.
He says 500 requests was gone fast but I see his requests for more than 30 minutes of work, he expects the simplest of all packages which I told him that was only suitable if you not heavy coding.
He expected to work many hours with Opus 4.8 on a 1M context and his context window was probably already extensive when he started because it jumped from 0 to 15 requests just to start and then stabilized. Its not a fresh session which will consume more and more requests because the AI when started on an extensive context it has to read history, all docs related to that session, system instructions , prompts and start writing code. Just these will consume a lot of requests.
I never ever made any promisses to these 2 guys and they expected something that was only in their heads.
My website footer and Docs has all the information they said I didn't provide. I don't have to tell every buyer every single rule if they don't ask about it!!
MY FINAL NOTICE: IF YOU DO NOT KNOW HOW THESE SERVICES WORK READ THE PROVIDER WEBSITE END TO END OR ASK QUESTIONS!! DONT JUST BUY IT AND THEN COMPLAIN ASKING FOR REFUND, THATS NOT HOW THE REAL WORLD WORKS!!
This post is by a banned member (VelvetV7) - Unhide
I went through your website and documentation AGAIN and still couldn’t find clear information about how tokens are calculated or how usage is actually billed. Below are the screenshots of the "information" that the seller provides.
Everything else is just guides on how to connect the API to whatever software you are using, so tell me how should I know? Where is all the "info" at?
The point you keep on repeating again and again is that people that used 3rd party services before already know, and understand, well why don't just specify in your thread/website that this service is only for people that used such services before? You don't want new customers? How is someone new supposed to know? Also the fact that you are soooo against refunds, even when you can check that the customers didn't use 95% of the tokens makes me feel like you are just reselling, since your whole website is literally 100% AI made, doesn't look professional AT ALL.
Also what is the problem with me only testing few prompts? I tested it out, it was billing me like crazy, ACTUALLY billing me TWICE per prompt, which you never explained why it does that, I didn't like it, and stopped, expecting you would refund me for the unused tokens like others would, but nope, you never give refunds, I wonder why, maybe because you are reselling making pennies, and can't afford to lose a single dollar.
So please tell me, where is all this information at? Where does it say how the tokens work, on official Claude/GPT and whatever other AI provider its stated clearly 1 token is around 3-4 characters, with your API a simple "Test" prompt was 35k tokens which I got DOUBLE BILLED for so 70k in total.
I am only talking for myself tho, I don't know what the other guy did, I just added my reply here, because I also had a bad experience with what you provide.
Also please can you finally let me know why is it constantly double billing as well? Or will you keep on dodging that, and just reply with "oThErS uNdErStAnD". Even before on telegram you told me that you never had a customer that didn't like your service, yet here we are another customer of yours that doesn't like what you are providing, strange.
This post is by a banned member (MrPoisonLab) - Unhide
I went through your website and documentation AGAIN and still couldn’t find clear information about how tokens are calculated or how usage is actually billed. Below are the screenshots of the "information" that the seller provides.
Everything else is just guides on how to connect the API to whatever software you are using, so tell me how should I know? Where is all the "info" at?
The point you keep on repeating again and again is that people that used 3rd party services before already know, and understand, well why don't just specify in your thread/website that this service is only for people that used such services before? You don't want new customers? How is someone new supposed to know? Also the fact that you are soooo against refunds, even when you can check that the customers didn't use 95% of the tokens makes me feel like you are just reselling, since your whole website is literally 100% AI made, doesn't look professional AT ALL.
Also what is the problem with me only testing few prompts? I tested it out, it was billing me like crazy, ACTUALLY billing me TWICE per prompt, which you never explained why it does that, I didn't like it, and stopped, expecting you would refund me for the unused tokens like others would, but nope, you never give refunds, I wonder why, maybe because you are reselling making pennies, and can't afford to lose a single dollar.
So please tell me, where is all this information at? Where does it say how the tokens work, on official Claude/GPT and whatever other AI provider its stated clearly 1 token is around 3-4 characters, with your API a simple "Test" prompt was 35k tokens which I got DOUBLE BILLED for so 70k in total.
I am only talking for myself tho, I don't know what the other guy did, I just added my reply here, because I also had a bad experience with what you provide.
Also please can you finally let me know why is it constantly double billing as well? Or will you keep on dodging that, and just reply with "oThErS uNdErStAnD". Even before on telegram you told me that you never had a customer that didn't like your service, yet here we are another customer of yours that doesn't like what you are providing, strange.
You were rude from the first message you sent after trying one single prompt, saying my service was useless and horrid, now you keep bringing your 0 knowledge here and asking questions from a point of view of someonee who doesn't know what they are talking about.
As I said if YOU don't know how a service works ask questions before buying this is rule number 1 ! I Provide my service and that's it, there's nothing else to declare here untill a mod comes to the talk.
![[Image: image.png]](https://i.ibb.co/GQ1dsQHZ/image.png)
![[Image: b.png]](https://i.ibb.co/VY4h7YhN/b.png)