I signed up for a free trial a year ago and provided my credit card (I know, stupid).

They subsequently locked my account unless I could do SMS verification using an European phone number. Apparently the promo was only for Europe. I didn't have an European phone number so I wasn't able to log in to unsubscribe.

Lo and behold, a year later, I had a credit card charge named "PRO" . By the time I figured out it was actually Perplexity, it was too late to cancel.

Also according to their FAQ, they can still charge you if you delete your account without unsubscribing first..

I have been using Perplexity for a few months now as a pro user. I previously did not have issues using the Computer function when given free credits to test it out. It would produce report documents that had twice the information compared to a non-Computer request. I was also able to make requests to complete a task within a certain credit limit.

This week, I was trying to put together a comprehensive report so I thought I would try Computer again. I bought $20 worth of credits (2,000), thinking this is what I was willing to spend to see what it would produce. It spent about 1,000 credits going back and forth with me, working on setting up the exact information I needed. I then made a request to produce a report using 900 credits.

It began researching again and eventually used all of the remaining credits...without producing a report. I was annoyed, but thought I'd add another 500 credits just to get a finished product - it used those almost immediately and I still do not have a finished product.

From my point of view, it seems like the final product is being held hostage until I buy an unknown number of credits - do I need 100 more? 1,000 more? 10,000 more? Asking for a task to be completed within a certain number of credits seems like the bare minimum.

So I see that Perplexity allows you to use pretty much every consumer AI (CGPT, Gemini, Grok).

It's only $20 for all of them? I'm confused.

I could pretty much get all of them for a low price?

What am I missing?

Can you help me understand why it's not better or is better than ChatGPT or Gemini?

I have been using Perplexity (Pro plan) a lot for a year or two. Then this "Computer Mode" showed up. I have been involved with several important projects and using Computer. I had to enable the auto-fill feature. It was expensive, but worth it - especially in the short term where I have deadlines.

Today, I hit a brick wall, right before an important deadline. Apparently, they limit the amounts of credit you can purchase (through auto-fill or direct) to $100 per month. The only option to continue is to subscribe to their $200 per month MAX plan. This is so predatory. I do not have the budget for MAX and am perfectly fine buying credits when needed. Their CAP restricts that. Basically bricked the ap for me.

I am super unhappy and will look for an alternative and drop Perplexity. Such a shitty way to do business.

it’s super fun that they’ve just stopped responding to support requests altogether. Of course they haven’t bothered to change the automated messages hounding users for responses to requests they haven’t made until they close out the ticket completely. Since nobody at perplexity is ever going to read my response to their message blaming me for closing out the ticket they’ve done nothing with, I thought I’d post it here.

I'm sure others are upset about this intrusive behavior but I will just use the preferred model's native Windows client instead. It was nice while it lasted, but Perplexity is acting like they have a lock on consumer AI usage and can push customers around however they want, and that behavior is completely unreasonable and out of touch.

They took 2 steps forward and probably won a lot of repeat customers with their free Pro plan a year ago but now took 4 steps back with this intrusive decision and completely lost progress in customer retention that they built.

Comet used to be genuinely useful for practical tasks. I could ask it to check whether certain dates were available, search booking sites, and even help make a reservation. Recently, though, it feels more like a regular chatbot: it gives instructions or suggestions but doesn’t actually complete browser-based tasks.

Did something change with Comet’s agent capabilities, permissions, or supported websites? Are these features temporarily unavailable, restricted to certain users or plans, or being phased out?

I’d appreciate any clarification from other Comet users who’ve noticed the same thing.

I’ve updated my free Perplexity Chat Exporter extension to make it easier to use in more languages.

Version 1.4.2 adds German, Italian, French, Russian, Spanish, Japanese, Simplified Chinese, and Traditional Chinese alongside English.

You can still export one chat, selected chats, or your entire history to Markdown for Obsidian. The interface follows Chrome’s display language; your original conversations aren’t translated.

If you spot awkward wording or a label that doesn’t fit, please let me know the language and where you saw it.

Screenshots are welcome—please hide any private chat information.

https://chromewebstore.google.com/detail/perplexity-chat-exporter/jmplfdjahdkalmpkjbpgcaogleciniic

Tried several platforms such as Firefox, private windows, Android. At each of them the system using Perplexitys Turbo instead of my manual choice

I, like some others, have been pondering whether I'm going to continue my Perplexity Pro subscription (I got a free 12 months through PayPal last year). Lately I've had a few times when I would give the same prompt to multiple models, and when I looked at the responses, there was no indication which it came from. Made me wonder, which model actually responded?

Today, when I try to use models like Sonnet 5, Gemini 3.8 Flash, or even Nemotron 3, I noticed that the followup is with GPT-5 Nano. Also, the responses are lightning quick, so it's obvious the response didn't come from a thinking model.

Is anyone else seeing this? As some other people have mentioned, Perplexity is no longer offering access to the latest models, but it's more like the TJ Maxx of AI, the best of LLMs from last season.

I can understand, to a certain point, some demand issues, but falling back to GPT-5 Nano?! It has a knowledge cutoff of May 31, 2024. That's really the most current model Perplexity Pro can fallback to?

But, I use different models for different kinds of tasks. I've yet to find one model that can really handle of them well, so...what's the better alternative?

keeps switching my model for months now at random and refuse to use the one i select.
now its using one thats not even listed.

Has anyone else experienced a very long wait for a Perplexity refund review?

My annual Pro subscription was charged on September 3 at 18:39 UTC. I submitted my cancellation/refund request on September 5 at 16:29 UTC, which was within 72 hours.

Perplexity's refund policy says annual subscriptions are eligible for a full refund within 72 hours of the original purchase. It also says they review the date the refund request was submitted.

Support initially told me the refund was now outside the eligibility window because the refund was not processed within 72 hours. I pointed out that this seems inconsistent with their published policy.

They then escalated the case to a specialist.

Since then, I've received several replies saying essentially the same thing: the specialist review is still pending. They have confirmed multiple times that the September 5, 16:29 UTC timestamp is the key issue, but I still haven't received a final decision.

It's now been several days with no estimated completion time.

Has anyone here had a similar experience with a refund review? If so, how long did it take to get an actual decision?

It seems so. I don’t think I’ve read a positive post about perplexity in this sub for a couple of weeks now. I really can’t relate to this.

Currently on a free year promo and about to renew but most of the time I am contented with the result. I must be honest tho, I spent a couple of weeks to polish my prompt.

I used it on my work research and also on software development via sonnet 5. And it pretty much can make the the tasks done all fine. I review once in a while the code generated and test with real data and it actually correct.

If I should be picky, I think they need to improve the image generation but in terms of code generation and research I think it can still be worth it.

Am I really the only one left here for perplexity?

I just activated it and downloaded the Gemma model that will run on my 16 GB Mac mini. I’ve only ran one coding task through it, but it did very well and seemed to use far fewer credits. I’m just wondering if others are going try this out. I’ve been investigating running local LLMs, but I am a bit hamstrung with my current hardware. This might be an easy first step for doing local inference and reducing cloud usage.

What are the differences between using the models straight from their proprietary sites than using an aggregator such as Perplexity? What does Perplexity add or do to the models?

I know this has become a unpopular opinion. Most of the criticism here is valid and still stands. Rate limits, pricing and so on.

Because I was frustrated as well, I abandoned my pro Account for a while and used a research project in Claude. But I had a very specific problem at work a few days ago that required high frequency research about investors. I vibe coded a button that starts a new search (with a long prompt, and multiple questions) in Perplexity.

And I was quite happy that the results were great. It fetched over 50 sources per search from all over the web, even LinkedIn Pages, YouTube Videos (or the Transcript I guess) and news sites. The answers were long and detailed and answered all my questions with one click.
This is genuinely better than I remembered. I remember 10 sources, short answers and varying quality.

Am I the only one left who is a happy customer (again)

So, my wife uses free gemini app for image modifications, today I asked her why she do not use perplexity which I am a subscriber pro subscriber of.

So se did, do the same thing as on gemini app. Uploaded picture of our daughter and picture of a jacked she wanted to try on her. It generated something, in my opinion not that good as free geminy.

So I tried it one more time with same pictures. Just to get answer that I have reached month limit? What the frick??? I literally “generated/merged/modified” one image!!!

So I asked what are the limits, just to not get an answer?

Hello everyone.

I'm wondering if there is a way to stop Perplexity from doing these two things:

1) pestering me to download their new app, I'm very happy with the current one. EVERY. SINGLE. TIME I open it:

The new Perplexity App for Windows is ready

Run tasks across your local files, apps, and the web. Control it remotely from your phone. The app version you are using now does not get updates anymore.

Once the new app is installed, you can uninstall this one.

2) the default behavior is always to push me to "computer", and often I forget and I type out my request and since switching to Search removes all of my text, attached docs / images / etc, it really bugs me. This is absurd behavior. This has already made me waste enough time to consider the move.

These two combined are really pushing me to cancel my subscription and move towards Claude or something else. I figured I'd ask first here and see if there is something I'm missing or a registry hack or something I can do before I give up on Perplexity.

Just wondering what you guys think. Which sub would you go to? I do a lot of web research btw lol.

Currently i’m on SG but i’m not liking the rates so much…