To me, Manus AI is the worst AI agent I have ever used, and I have used many because testing AI tools is part of what I do. I have used enough of these products to know the difference between something rough and something that should not be taking customer money. Manus is the one that stands out the most because it is the most incomplete paid AI product I have used.
Manus can occasionally do something right, which is one reason people keep giving it another chance after it fails. That does not change how often it gets simple things wrong, ignores instructions, wastes credits, breaks progress, and creates problems the customer never asked for. A product can be bad on its own, but it becomes a different issue when billing, refunds, credits, and support are just as unreliable.
I already did not trust Manus before the duplicate charge happened. I had already dealt with failed tasks, wasted credits, ignored instructions, and damage to real work. The duplicate charge made everything worse because customer money was now involved outside the normal credit system. Manus charged me twice, acknowledged the duplicate billing issue, and still has not refunded me.
I also still have not spoken to a real human from Manus about the duplicate charges. That matters because the issue is not hard to understand, and it does not require an agent session, coding review, or technical investigation. I reported duplicate billing, Manus support acknowledged it, and Manus still has not returned the money. That is not a complicated AI problem, and Manus should stop treating it like one.
I contacted Manus support and asked them to remove the duplicate charge. I was not asking Manus to debug an app, review code, inspect a failed project, or understand some unclear complaint. I was asking Manus to refund duplicate billing on my account. That should be one of the easiest problems for a paid software company to correct.
Jordan joined the conversation and identified itself as an AI agent from the Manus support team. Jordan understood what I was reporting, said it could see the duplicate charge, looked into the account, and sent the case to billing for further review. At that point, Manus already had the account, the charge information, the duplicate billing report, and the refund request. I should not have had to keep chasing the same issue after that.
A real billing process should have corrected the duplicate charge and confirmed the refund. Manus did not need me to explain what duplicate billing means, and it did not need another support bot to rediscover the same issue later. The duplicate charge had already been reported, acknowledged, and sent to billing. Instead of a refund confirmation, the issue came back through another support response.
Maria repeated that I wanted a monetary refund for the duplicate charge, even though Manus already knew that. Jordan had already acknowledged the issue and said billing would review it. Instead of confirming that the duplicate charge had been refunded, Maria said the case needed to be transferred to a specialized internal team for further review. Manus also told me unacknowledged cases could be closed after 24 hours, even though Manus had already taken nearly two weeks to return to the same issue.
That is backwards support, and it is not how a company should handle customer money. Manus delayed a basic billing issue, then put pressure on me to acknowledge another transfer so the case would not be closed. A duplicate charge is not a coding bug, hallucination, task failure, product limitation, or missing prompt detail. It is a billing problem, and Manus should not need multiple AI support bots to understand whether a customer was charged twice.
I pushed back because Manus had already made me distrust its refund process. I previously asked for a monetary refund after Manus damaged my work, and support told me a refund had been issued. The money did not arrive, and that changed how I read every refund promise from Manus after that. When a company says a refund was issued, the refund should show up without another support fight.
That is why I do not trust vague refund language from Manus anymore. Billing will review it does not mean the refund will arrive. A specialized internal team will review it does not mean the money is coming back. Support saying something has been escalated does not mean anything has actually been fixed.
The duplicate charge is the center of my issue because it is happening to me now. Manus still has not refunded the duplicate charges, and Manus still has not connected me with a real human about them. The amount is not what matters here, and I am not publishing the amount because the amount is not the story. The story is that Manus acknowledged a duplicate billing issue and still has not resolved it.
The same support pattern showed up with a separate credit refund request. I contacted Manus after a task session wasted credits and made an unauthorized destructive change to a live production product. I gave Manus the task share link and explained that the agent ignored explicit credit budget instructions, replaced a live production component without permission, installed and removed an unnecessary package, wasted credits on repeated messages and dead-end commands, failed to fix the original issues, and required a rollback to undo the damage.
That request was not vague, and I did not leave Manus guessing about what happened. I gave Manus the task link, the component issue, the field problem, the mapping problem, the unauthorized replacement, the package mistake, the rollback checkpoint, and the original problems that were not fixed. I was asking Manus to return credits its own agent wasted while making the task worse. I was not asking support to figure out the situation from nothing.
Jordan joined that credit refund conversation too and again identified itself as an AI agent from Manus support. Jordan said it could see the problem, reviewed the session, and forwarded the credit refund request to the relevant team for handling. That sounds official, but it still left the customer waiting while the issue moved away from a real resolution. The same company that wasted the credits was now putting the refund request into another support lane.
Then more Manus support bots joined and the same behavior continued. Chris transferred the case again, Mel asked for screenshots, error messages, video recordings, or more material even though the task link and full explanation were already in the chat, and Kench later asked for a screenshot before escalating the request to Manus’s Specialist Team. The issue did not become a clear credit restoration, refund confirmation, billing correction, or payment record. It became another support loop.
That is the problem with Manus support in my experience. The names change, but the behavior stays the same. The AI support bots repeat the issue, ask for information already provided, transfer the case, pause the chat, or send the same specialist team language. The customer still sits there without the one thing that matters, which is the money back, the credits restored, or proof that the issue was fixed.
Manus is using AI support bots to handle complaints about an AI product that already wasted credits, made unauthorized changes, failed to deliver the work, and left the customer asking for money back. That is exactly why the support process feels insulting. I am dealing with a broken AI product, then I am forced to deal with more AI when I ask the company to fix the money side of the damage. The same kind of system that caused the problem is standing between me and a refund.
The public review profile makes this look even worse. Based on the Trustpilot numbers I saved, Manus showed a lower TrustScore than reservecreditcard.com, which is a site I already view as a scam-style operation because of complaints involving fees, refunds, delivery, payment, and customer service. Manus being rated lower while people complain about payments, refunds, credits, subscriptions, and support makes the customer trust problem harder to ignore. I am not relying on Trustpilot to tell me what happened to me, but the comparison matches the kind of money problem I am dealing with.
I believe Manus has a refund problem because Manus charged me twice, acknowledged the duplicate billing issue, still has not refunded me, and still has not connected me with a real human about it. The public complaints only make my own experience look less isolated because the same subjects keep coming up around money, support, credits, and refunds. A paid AI company should not have a worse public review profile than sites people already warn each other about. That should make any customer think twice before trusting Manus with a card.
Manus should not be taking customer money like a serious software company while handling refunds like this. A serious company corrects duplicate charges quickly, confirms refunds clearly, restores credits when its own product wastes them, provides real records, and does not make customers keep proving the same issue after support has already acknowledged it. Manus has not done that for me. Manus still has not refunded the duplicate charges.
Until Manus refunds the duplicate charges, confirms monetary refunds properly, restores wasted credits without repeated AI support loops, and stops hiding billing and task failure issues behind specialist team language, I would not trust Manus with my money.
- Cloudflare Says Anthropic Mythos Can Chain Bugs Into Working Exploits
- Manus AI Review: The Worst AI Agent I Have Ever Used
- How Claude Deleted the PocketOS Database in 9 Seconds
- Google Commits Up to $40 Billion to Anthropic While It Still Tries to Sell Gemini
- Anthropic’s Claude Mythos Fell Into the Wrong Hands
Sean Doyle
Sean is a tech author and security researcher with more than 20 years of experience in cybersecurity, privacy, malware analysis, analytics, and online marketing. He focuses on clear reporting, deep technical investigation, and practical guidance that helps readers stay safe in a fast-moving digital landscape. His work continues to appear in respected publications, including articles written for Private Internet Access. Through Botcrawl and his ongoing cybersecurity coverage, Sean provides trusted insights on data breaches, malware threats, and online safety for individuals and businesses worldwide.










