Breach of Contract – Unlimited Plan Downgraded Mid-Subscription
Myrddin
Hello Merlin team,
I'm writing to formally raise a serious complaint regarding recent changes to the Merlin yearly plan I purchased.
I bought the Pro plan for a year, which you advertised as having unlimited credits.
Unilateral Changes After Sale
I purchased the Pro plan as a yearly subscription with unlimited credits under a fair use policy. Around April this year, you abandoned this promise and suddenly introduced credit limits for existing Pro plan users. Changing the usage limits after you sold the deal is a breach of contract. You cannot revise the terms of a deal after both parties have agreed and payment has been made.
Unworkable Daily Limits
Not only did you introduce credit limits, but you also imposed a daily limit by dividing a monthly cap into 30 or 31 equal parts. This method does not reflect real-world usage. Some days I use the tool more, some days not at all. Enforcing a strict daily quota makes the tool impractical and restricts fair use as originally promised. If a monthly limit is imposed, it should be applied monthly—not daily.
Tool No Longer Usable as Intended
Under the previous policy, I was able to use Merlin for all my daily tasks. Today, after asking just two simple questions to Grok 3 (listed as 25 tokens per question), I’ve already used up 9% of the daily allowance. Based on this, it seems the daily cap is roughly 550 tokens. This has nothing to do with a “fair use policy.” It is an artificial cap that severely restricts functionality and makes Merlin practically unusable for daily use.
I fully understand the need to combat abuse—this is exactly what a fair use policy is meant for. But this current system is not about preventing abuse; it simply limits regular use.
In Summary
You marketed and sold a product with unlimited tokens, then reduced those capabilities mid-subscription, and are now suggesting users pay again for features they already had. You even removed the “unlimited” label from the user interface. This is not acceptable and constitutes a breach of the original agreement.
I already contacted support about this issue, but my concern was dismissed. I was simply told to pay more if I wanted to top up my credits. This completely ignores the fact that the original deal has been altered without consent.
I expect Merlin to honor the terms of the subscription as sold—nothing more, nothing less.
Vijay Bharadwaj
(1/n) Hi everyone,
I've read through all your concerns, and I want to address each one personally and thoughtfully.
First and foremost, I genuinely apologize to anyone whose experience has been impacted by how our product has been designed or communicated. This comes from my heart -- your satisfaction matters deeply to us.
Before diving into specifics, I'd love to share our perspective, and I hope you'll read this with an open mind, just as we're listening to yours with open hearts.
Let me address your concerns:
- Fair Usage Limits: Daily and monthly limits have been part of our platform since we introduced the Pro plan. I completely understand how frustrating limits can feel when you're in the flow of using AI. Recently, after hearing your feedback about restrictive limits, we actually increased daily limits by 50% for Pro users while maintaining monthly limits.
These limits serve two purposes: ensuring sustainable operations and preventing system abuse. While I truly empathize with how this might feel limiting at times, these measures help us keep the service reliable and accessible for everyone, and this is not something to be taken lightly.
- Usage Running Out Quickly: We hear you on this, and it's why we proactively added a transparency banner at the time of launch, explaining that our new agentic chat features incur more costs due to the complexity of tool-calling at scale.
For context, a web search that previously cost 2x ⚡ now costs 3x – but the results speak for themselves. Merlin Search (even with something as cheap as Grok 3 Mini) delivers results that outperform Perplexity Pro Search anyday (which is surprisingly considered gold standard in AI search). Our team is working tirelessly to optimize efficiency, but we believe in honest pricing, not undercharging to save our asses.
That being said, we're soon launching a High Efficiency Mode that will be significantly more cost-effective. It will have some trade-offs in context window, response quality, and memory size, but it gives you more choice in how you use our platform.
We recognize that different users have different needs. If your primary use case is asking many quick questions without needing advanced agentic capabilities, other products might better suit those specific needs -- and that's okay! Our vision is to be the best general-purpose AI agent, and we believe we're making meaningful progress toward that goal.
n
ninix23284@gotemv.com
They deleted this btw
Merlin
ninix23284@gotemv.com: Hi Ninix, That post is essentially highlighting the similar point mentioned in this post. Please refer to the comments of this discussion and keep the discussion as part of this thread. Thank you
Vijay Bharadwaj
(1/n) Hi everyone,
I've read through all your concerns, and I want to address each one personally and thoughtfully.
First and foremost, I genuinely apologize to anyone whose experience has been impacted by how our product has been designed or communicated. This comes from my heart -- your satisfaction matters deeply to us.
Before diving into specifics, I'd love to share our perspective, and I hope you'll read this with an open mind, just as we're listening to yours with open hearts.
Let me address your concerns:
- Fair Usage Limits: Daily and monthly limits have been part of our platform since we introduced the Pro plan. I completely understand how frustrating limits can feel when you're in the flow of using AI. Recently, after hearing your feedback about restrictive limits, we actually increased daily limits by 50% for Pro users while maintaining monthly limits.
These limits serve two purposes: ensuring sustainable operations and preventing system abuse. While I truly empathize with how this might feel limiting at times, these measures help us keep the service reliable and accessible for everyone, and this is not something to be taken lightly.
- Usage Running Out Quickly: We hear you on this, and it's why we proactively added a transparency banner at the time of launch, explaining that our new agentic chat features incur more costs due to the complexity of tool-calling at scale.
For context, a web search that previously cost 2x ⚡ now costs 3x – but the results speak for themselves. Merlin Search (even with something as cheap as Grok 3 Mini) delivers results that outperform Perplexity Pro Search anyday (which is surprisingly considered gold standard in AI search). Our team is working tirelessly to optimize efficiency, but we believe in honest pricing, not undercharging to save our asses.
That being said, we're soon launching a High Efficiency Mode that will be significantly more cost-effective. It will have some trade-offs in context window, response quality, and memory size, but it gives you more choice in how you use our platform.
We recognize that different users have different needs. If your primary use case is asking many quick questions without needing advanced agentic capabilities, other products might better suit those specific needs -- and that's okay! Our vision is to be the best general-purpose AI agent, and we believe we're making meaningful progress toward that goal.
Vijay Bharadwaj
(2/n) 3. Contract Changes: I want to be crystal clear: there have been no contract changes. The limits have always existed, communicated and we've only worked to make them more transparent on our website. Pro was NEVER marketed as "unlimited usage regardless of model or usage pattern" -- and here's why:
You could technically use Gemini 2.0 Flash extensively and achieve what feels like unlimited usage. But would that justify claiming "unlimited"? Some might say yes, because you can ask unlimited questions and get unlimited answers. (After all, answers are what hold value, not tokens.) Others would strongly disagree... and I see both perspectives.
The reality is that the entire AI industry faces this challenge. The technicalities of how LLMs are distributed by providers (OpenAI, Anthropic, etc.) makes it incredibly difficult to create a truly unlimited service that's both sustainable for businesses and satisfactory for users. If you check forums for Cursor, ChatGPT, Perplexity, or Manus.ai, you'll find similar discussions – but at 10x the volume. This isn't an excuse, but it's context for an industry-wide challenge we're all navigating together.
In short, give people better? You give them less. Give them more? You give them worse. You give them more AND better? You die within a year.
What makes us different (and why our users appreciate us):
Transparency: We show your current usage right in the product
Clarity: We clearly state what usage you receive
Value: You get 4-5x worth of AI usage compared to what you pay ($20/month gets you $100/month worth of usage)
I genuinely believe this approach is generous, respectful, and represents a better way to run an AI business – especially when AI has become such a crucial tool for work and life.
Please know that this message comes from a place of deep respect for our community. We're a small, passionate team trying to do right by our users while building something sustainable and innovative. Sometimes we might not get everything perfect, but our hearts are in the right place. We'd rather pour our energy into creating breakthrough AI experiences than constantly tweaking pricing models.
Therefore, please understand, and do not make this a David v Goliath duel. We're as David as you all are. I apologize that sometimes we can't give you more than "we're working on it", but what I said above is everything that goes unsaid when we say "we're working on it".
Vijay Bharadwaj
(3/n) P.S. About NSFW generation: This is a sticky spot for us, because we have to tread extremely carefully with our AI providers to ensure we don't get banned. This is, unfortunately, strict policy across all models on Merlin, and we will not budge on this. But, fortunately, you have a ton of other products to choose from, and again, that's okay.
闪亮闪亮
First they limit the conversation on "Pornography" chats, pottentially to exclude a large base of users that using it for writing some novels and some NSFW chats. Just a little trigger would straight cut out the conversation and will no longer generate anything.
Then they figure that isnt enough with so many users doing different thing, then they decide to put limitations here and there to excludes another big chunk of customer.
And third they are moving on to this "fair usage" policy. Btw I am not going to renew with them and trying to fends off new customer that pottentially wants to buy from them.
S
Siddhartha
Hey Folks, first of all, we are sorry for all the inconveniences that you all are facing. We understand the pain, and its been hard for us too. We want to give unlimited everything, we ourselves want to use AI in that way. But as we are trying to make Merlin more and more powerful, with better models (all top models are thinking models now). AI has gone through a curve that things are more expensive now verses 6 months back.
The fair use limits that were untouchable 6 months back, are now much more reachable due to better models and agentic chat, that tries to understand userstand user query and answer that. We have pushed a change to increase the daily fair use limits, and we hope that should alleviate some of the issues. We are always trying to give the best service possible at all times and it gives us no joy when users face issues in their day to day work.
n
ninix23284@gotemv.com
Siddhartha that wouldn't change a thing since the monthly limit still the same, it makes user consume more than they should have daily and monthly
Please just remove the monthly limit and keep the daily limit
That keeps the unlimited queries and fair usage limits
Or better yet, just make Gemini pro, deepseek, grok mini as unlimited for pro users even if you keep the monthly limit, as that consume less
Or any 40 and below credits as unlimited (make it still consume daily limit but not the monthly limit)
S
Simon
endu Why? Why? Why???? Why are you moved on this way to communicate with your customers.
First the the big desaster with the discounted PRO Plans where you just took the "Pro" from plans - afterwards (yes you mentioned in th contract that you are allowed to do... but month after we see, this is more a common business move from your side).
the daily limits / monthly limits are fully unclear - for everyone.
Also the way how you communicate new update massivly .. from here on the feedback webpage, to the inapp notification, currently no idea how you communicate it.
IF YOU DO NOT WANT THAT CUSTOMER USE THIS TOOL - that is oky .. but just be more open here. There are many other competitors which maybe has another focus in product strategie, but also another mindset how to handle customers.
sorry, you are nearly to loose one of a real merlinAI fanboy over the last month... sorry to say. Hope you will come back on track within the next weeks, otherwise I will leave merlin and not expand.
I also expect Merlin to honor the terms of the subscription as sold—nothing more, nothing less!!!
Please stop ignoring the customers voice!
Best regards
t
theunmindful
endu You may review the credit system which is transparent. It creates a sense of deficiency and usually seen negative by the users (as studies show). This is the reason why Sora credit based usage was removed.
An alternate pragmatic approach would be what Merlin does the best.
Rather than a fixed daily quota, keep it dynamic based on the collective usage by the users (backend). Limiting models and rationing especially expensive models during cumulative high traffic. You may stay transparent about this collective traffic (green yellow amber) so that users know the status of usage. Of course you can keep the upper limit for daily and monthly (to check for abuse).
You have sufficient user data to analyze it and find the sweet spot. A win-win.
Y
Yume
theunmindful Agreed 1000%
n
ninix23284@gotemv.com
At the very least, unlimited plan in fair use should be similar to simply have a daily limit, but remove the monthly limit.
It restrict user to fair use while following the contract of "unlimited queries"
Y
Yume
ninix23284@gotemv.com Disagree because some users may use heavily more one day than another so monthly caps based on plan (okay sure) but the point is and still remains that's not what a lot of us were "sold" so I agree with @myyrdin where this can be construed technically as a breach. Companies do NOT switch plans after purchase as many times as this company has... Let's keep it real here.
Kind of like the old "bait and switch"
T
Taz
I concur. I complained via support ticket and was provided with a wishy washy explanation and a very confusing explanation about their token use that has nothing to do with me as a customer. I no longer get what I paid for and most of the time resort back to chatgpt. Thanks for posting this. It reminded me to write a very honest review on a review site. Totally unacceptable of this company.
Y
Yume
Taz Oh if I shared the screenshots of recent responses of one of support it's like no clear answer no clear answer no clear answer game of Twister to try to figure out what that guy was saying. Ironically then he says, oh by the way your plan is up for renewal end of Aug. Seriously as if tone deaf was a person...
There's definitely a disconnect going on with support.
Baptiste D.
Whilst I think MerlinAI has come a long way in terms of usability and practicality, I do agree with you. If i go to my subscription setting, it does say I am a pro user with unlimited queries which is no longer the case.
The usual defense is, it costs a lot, we still need to make a profile and read our terms. Whilst yes, you should read the terms when signing up with such things as AI aggregates, it does feel like i eat up through "credit" or your fair use limit quicker than before.
If you guys reject adding more varied models to the service, onky stick with big names but then lower the user's usage, how is the user winning? Other services like Merlin, which I won't name offer a more detailed and comprehensive approach to what usage limits is, whilst we are still stuck with a percentage bar.
For the price, i do want to consider elsewhere, and the thread OP does make a point saying thr whole usage thing changed mid contract without warning to the users.
Please be more open. Stop telling us you are working on it and making it clearer. Make a community post, or send an email out where you explain the new terms of use. If there are inequalities between plans (as you said in the past, the cheaper or promotional pro plans have lower limits but are still called pro) state them. Change the progress bar. As a pro user you can see credit cost to send a message to a LLM, but then it gets converted to percentages.
Your product is good. The UI is good, but the terms are still hazy to even long term users and it feels you guys have been using the "we are working on it" excuse for a little bit too long.
Reda Izo
so i'm not crazy, i went from 1500 credits to 200 credits in one evening by itself. Wahts happening ?
Load More
→