Am I the only one who feels that existing users who actually paid for credits and put their faith in ParakeetAI are increasingly being sidelined, while new users continue to receive free sessions to attract them to the platform?
Let me explain why I feel this way.
ParakeetAI seems to be changing or updating the product very frequently its models, features, response behaviour, limits, and other aspects of the experience can change from one version to another. As a paying customer, shouldn't we have at least a small opportunity to test these changes before relying on them in an actual interview?
For example, if the product is changing frequently, why can't existing credit buyers receive something as reasonable as 10 free practice sessions per release/update to evaluate the current version?
I'm not asking for unlimited free usage. I'm talking about a limited opportunity to answer very practical questions:
Is the new model fast enough for a live interview?
Has response latency changed?
Does it follow my context and instructions as well as before?
Have recent updates improved or degraded the experience?
Do I need to switch models before my actual interview?
This becomes even more important when the interface itself can indicate that a particular model may be “very slow.” How is a paying user supposed to determine whether that model is actually usable in a live interview without testing it first especially when testing consumes purchased credits?
What concerns me most is the apparent difference in treatment between prospective users and existing paying users. New users can receive free sessions to evaluate the product, while customers who have already purchased credits can potentially be told that their historical use permanently disqualifies them from receiving even a limited number of sessions to evaluate the current version.
And if purchasing additional credits can somehow restore access to those free sessions, that raises an even bigger question: why should an existing customer have to spend additional money merely to regain the ability to test the product?
I also want to acknowledge that Claude Sonnet 4.5 has now appeared again in the latest update, which is welcome. But the fact that it is available again doesn't solve the practical issue: if the system indicates that the model may be very slow, I still need to test its actual response time before using it in an interview.
Personally, I also feel that some of the earlier versions of ParakeetAI were considerably more user-friendly not just in terms of the product itself, but also in how the free-session and usage policies worked. The newer approach increasingly feels focused on controlling usage and optimizing the commercial side of the service rather than giving paying users the flexibility they need to actually use the product effectively.
Perhaps I'm wrong. Perhaps others have had a completely different experience.
So I'm genuinely curious:
Am I the only one who feels this way?
If you are a ParakeetAI user (particularly a paying/credit-based user) I'd appreciate your perspective. Upvote, downvote, or leave a comment. I'm interested in knowing whether this is simply my individual experience or whether other users also feel that existing customers are being overlooked.
At the very least, I hope feedback from paying users encourages ParakeetAI to reconsider how frequently significant changes are introduced and, when they are introduced, provide existing credit buyers with a reasonable opportunity, even something as simple as 10 free sessions to test and understand what has changed before they have to spend their purchased credits.
Products like this ultimately exist to help people perform better. It would be unfortunate if, somewhere along the way, the original user-focused purpose gets lost in the pressure to optimize for revenue.
Sometimes businesses need their customers to remind them of the purpose they started with in the first place.