Claude Reviews 1,296

I've been double-billed CA$648.66 since March 22, 2026. Error 15WQOCG during Team workspace creation caused a duplicate workspace to be created and billed. I cancelled the duplicate immediately, but it's STILL being billed 30 days later.

In 30 days, I have sent 5 emails to support@ and received ZERO human responses. The only response came from "Fin AI Agent" (their bot), which told me in the SAME message that I'm "not eligible for a refund" AND that "a refund has already been provided." These two statements are mutually exclusive. When I pointed out the contradiction, the bot immediately closed my case.

There is no phone number, no live chat, no way to reach a human being. The support system is entirely automated and clearly broken.

I'm a business customer in Quebec, Canada. Billing for cancelled subscriptions violates Quebec's Consumer Protection Act. I'm filing formal complaints with the OPC and AMF.

If you're considering Anthropic for your business, know that if anything goes wrong with billing, you will have absolutely no recourse. Thirty days and counting.

Claude, when it works, can be useful. But it often does not work. It is down often. The recent upgrade caused authentication issues that now prevent me from accessing Claude.ai entirely. And there is NO customer service for non-enterprise customers. Contacting support(at)anthropic.com connected me to the Fin AI Agent. The AI agent’s solution was for me to contact support(at)anthropic.com. I had no choice but to cancel my subscription.

DONT USE CLAUDE IF YOU WISH TO FINSIH WORK ON TIME!!!! WASTE OF TOKENS, MONEY AND ESPECIALLY TIME, YOU WILL GET BETTER RESPOSNES FROM FREE CHATGPT, CLAUDE WILL WASTE TIME (MAKING YOU FEEL SOMETHING IS ABOUT TO HAPPEN GOOD) AND THEN DO EVERYTHING BUT NOT WHAT ASKED FOR AND KEEP HALLUCINATING

I subscribed to a monthly membership for Claude 'Sonnet' translation service. I started on a project commissioned by one of my customers.
2 weeks into the project, Claude was unavailable on line due to 'technical difficulties'. This access continued for another 2 weeks. During that period the company charged me automatically for the month - no service, but still charged.
I have requested a cancellation and a refund to NO AVAIL.
The company has no customer support other through an automatic 'BOT'. There is no HUMAN contact.
I am not only disappointed, but angry with the lack of service. I have attempted to resolve this matter through the web site using messaging and email but there has been no response.
I will lodge a complaint with Consumer Affairs in Australia.
Alain.

I tried Claude today and was impressed, at first. Then it kept making mistakes and redoing itself so i hit a limit. Then it says 'UPGRADE TO PRO!. So I did. After upgrading to Pro I WAS STILL AT MY LIMIT!!!
What a piece of trash. If you're serious about using Ai, stick with ChatGPT; I've gone hours and hours without a limit. Thanks for the refund Claude.

I tried Pro Plan, but it sucks totally. Only burning weekly tokens and hallucinating non-stop generates crap code. Not recommend. Stupid and useless, full of only unnecessary plugins and crap no one needed.

Horrible. Broke my code while running multiple agents with unrelated tasks -> claude breaks the code -> I don't want to revert because other agents made additions too -> I ask Claude to revert -> in the middle of reverting it runs out of the X hourly quota nonesense -> leaves me broken codebase. Horrible, by far the worst LLM I have used thus far. Unacceptable really. And then later it of course didn't have the full context anymore and couldn't repair it. Never again Claude. You can barely ask it anything or run into limits, which I also haven't at all experienced with any other LLM. Way slower for sometimes a stronger answer for small code solutions, but garbage as soon as app size grows.

This service is terrible, the worst AI out there. It does not listen to what you ask or remember anything you say in earlier chats. Above all, even when you pay for the pro subscription you only get 7-8 messages before they want MORE money from you. Good riddance hope they get banned.

I gave Claude a second chance because the free tier initially impressed me enough to subscribe. Unfortunately, the paid experience was worse, not better.

I spent my entire 5-hour usage window trying to get one prompt through. Responses failed, were canceled, or forced me to retry and resend the same prompt over and over. In the end, I got nothing done.

What makes this especially bad is that these failed and canceled responses still burned through my limited usage, leaving me stuck waiting another 5 hours just to try again. For a paid subscription, that is awful.

Support only made it worse. I was pushed into AI support loops, and my case was closed again without any real resolution.

Based on my experience, the subscription was not worth it. Paying for access only to waste an entire usage window on failed responses and then being forced to wait hours to try again is unacceptable.

After paying for packages, price problems arise and remain unsolved, with the money simply being stolen.
Furthermore, there’s no human support available and the AI bot follows a completely paperclip method. It’s horrible.

If anyone takes their business seriously, they should never use AI for replies; otherwise, it will be shut down soon.

do not buy this app if you have an alternative. I am trying GOOGLE STUDIO AI but it sucks still. No wonder Claude's reviews are so low. They do not respond ever to your help requests. They shut down any new options for new requests, so I came here. It is VERY obvious they could give a 💩 about the feedback here or anywhere. I sincerely hope I don't need to use this 💩 show app.

They recently changed the usage allowance without telling anyone and burned ALL my tokens. The system got stuck for 8 hours and burned all the tokens I just bought

we are far from a good system... they continue limit your things day by day, not week by week..

code not work great and also are full of useless code..

in specific time of day not work in any way, keep 20 minutes to think to a request then tell it was not able to reply.... the limit in chat messages are absurde.

but they want some money of big corp as grok, chatgpt... there are many solutions way better

It has been three days trying to get one article proofread. I provided the text and background material, and it still hasn’t finished a single pass. It hit limits three times on the same article without saving anything, freezing midway and saying I’d run out of allowance while it was still processing. The article is only 12,000 words—nothing unusual. DeepSeek has done it five times for free; this couldn’t do it once with a paid subscription. This isn’t a product; it feels like a scam—an appalling way to extract money. I won’t renew the subscription or recommend it to anyone.

only works the first 2-3 days, can get 20-30 tasks done, once you start using extra usage it gets 1 task done for every $5-10 added
nerfed

Was spending $600 on three accounts, built company workflows around this, integrated into my daily life. They regress old models breaking workflows and then released an unusable model. Will not be making that mistake again. Do not trust this company with your business, your workflows, or anything that you need to rely on.

Hi,

I'm writing with a tip about a situation involving Anthropic that I believe may be worth reporting. The short version: Anthropic's own backend bug caused a chain of events that led to my account being suspended — and after 9 days, the account remains inaccessible despite Anthropic confirming in writing both the bug and the subscription restoration.

I have documentation for every step.

────────────────────────────────
WHAT HAPPENED
────────────────────────────────

April 10 — I redeemed an official Claude Max 20x gift
card ($200/month plan, 100% covered). Anthropic issued
a receipt confirming the subscription through May 10.
(Invoice SSXODILM-0001)

April 12 — A backend bug in Anthropic's system
incorrectly downgraded my subscription two days after
activation, instead of maintaining the Max 20x tier.
I lost access to the plan I had paid for.

April 15 — After days of no response from support,
I attempted to restore access by adding a credit card.
The card lacked sufficient funds; the charge failed.
This triggered Anthropic's automated risk system and
my account was suspended.

April 16 — Anthropic sent me a written email
confirming two things:
1. The gift card subscription has been restored
2. The cause was a backend system bug on their end

April 19 — My account is still suspended. The
subscription is confirmed as valid and restored.
The account is still locked. I cannot access the
service at all. Nine days of a 30-day paid
subscription have been lost.

────────────────────────────────
WHY THIS IS A STORY
────────────────────────────────

This case raises several questions that may be of
broader public interest:

1. ANTHROPIC CONFIRMED THE BUG IN WRITING —
YET THE ACCOUNT REMAINS SUSPENDED
Anthropic's own April 16 email acknowledges the
bug and confirms the subscription is restored.
Despite this, the account suspension — which
exists solely because I responded to their bug —
has not been lifted. Anthropic has effectively
acknowledged fault while allowing the consequences
of that fault to remain in place.

2. A PAYING CUSTOMER SUSPENDED FOR RESPONDING
TO ANTHROPIC'S OWN ERROR
The failed credit card charge that triggered the
suspension only happened because Anthropic's bug
incorrectly downgraded my subscription. Without
the bug, there would have been no downgrade, no
credit card attempt, and no suspension. The
automated risk system penalized a customer for
reacting to a platform failure.

3. COMPLETE BREAKDOWN IN SUPPORT RESPONSIVENESS
Despite multiple support tickets, formal appeals,
a BBB complaint, and an FTC complaint, I have
received no substantive resolution in 9 days.
Anthropic's support infrastructure appears unable
to coordinate between the team that restored the
subscription and the team responsible for lifting
the account suspension.

4. GIFT CARD SUBSCRIBERS MAY BE SYSTEMATICALLY
AFFECTED
Anthropic has been distributing Claude Max gift
cards as promotional items. If the bug that
affected my account is not isolated, other gift
card recipients may be experiencing similar
downgrades without knowing the cause or how
to escalate effectively.

────────────────────────────────
DOCUMENTATION I CAN PROVIDE
────────────────────────────────

• Official Anthropic receipt (Invoice SSXODILM-0001)
confirming Max 20x subscription, April 10–May 10

• Anthropic's April 16 written confirmation
acknowledging the backend bug and subscription
restoration

• Screenshots of the account suspension notice
and "Missing Permissions" error

• Full support correspondence showing 9 days
of unanswered appeals

• BBB complaint confirmation

• FTC complaint reference

I am willing to speak on or off the record, share
all documentation, and answer any follow-up questions.
If this is something you'd like to pursue, please
feel free to reach out.

Thank you for your time.

After sending Claude 2-3 messages, the system makes me wait 5 hours. I paid for the Pro version under very difficult conditions. This kind of problem isn't explicitly stated on the purchase screen or anywhere else. This isn't a marketing strategy; it's fraud. Whoever suggested this marketing approach should be fired; it would be a significant positive step for Claude's prestige and future.

the fist day is great but as over time the ai model lets down on you with the terrible way of telling you to get there pro plan and if not you have just use your chat limit for the day. Even gemini lets you continue with there fast plan claude just does not do that.

This tool is not the problem — the company behind it is.

Right now, Claude AI feels like a victim of its own owners. If they don’t change their policies, they are going to destroy this product themselves.

Let’s start with the biggest issue: there is almost no real difference between the free plan and the Pro plan.
You subscribe expecting something better — and you get almost nothing extra. It feels like a trap: they push you to upgrade, but don’t deliver a better service.

Then come the limits — and this is where things become unacceptable.

There are limits everywhere:

limits per hour
limits per week
limits on text uploads
limits on image uploads
limits on conversation length

And these are not reasonable limits. They are extreme.

After a short time, your chat is cut, your workflow is broken, and you can’t continue properly. You simply cannot do serious work like this.

I use tools like ChatGPT, DeepSeek, Perplexity AI, and SciSpace — including their paid plans — and none of them are this restrictive. Not even close.

What makes it worse is the feeling of being locked in.
You start a project, and then you’re stuck. You can’t easily move your work, so you’re forced to continue — not because the tool is great, but because leaving is inconvenient.

Honestly, if I hadn’t already started a project with Claude, I would have cancelled my subscription immediately.

And the price? Way too high for what you actually get.

The frustrating part is that Claude is not a bad tool. It has real strengths, especially in coding. But all of that disappears behind these terrible policies.

This is not a technical failure — it’s a management failure.

I gave it 2 stars, and that’s only to be fair. It could have been much better than this.