We have a concurrency limit of 250 and are paying USD 12,000 per month, yet we are not receiving any support. There are no responses to our emails, and there is no assistance from the team. Additionally, the API does not allow more than 5 calls within 10 seconds. This is extremely disappointing and reflects very poor service.

Hi @Prio

We need ugent support

Can you send your email

We are not able runs calls

From which you mailed

support@cymake.com

So what's your org id

org_EebxiLykqLoay2Qx

Ok I send this to team to look

I need update urgent please

Yes i will mark this as high priority

It was five few days back

So I get back to you with the answer

ANy update ? @Shehroze -Support @Retell

I have reported the issue and when I will get the update I will tell you

frankly, the way your team keeps changing behavior without notice is becoming impossible to work with. Every time there’s no email, every day new minor update, or some undocumented change on your side, things break. No warnings. No proper changelog. No stability. This is not how a serious platform operates. Let me be very clear............ This level of inconsistency makes it extremely hard to survive in production. We are building real systems for real clients, not test projects that can tolerate random throttling and surprise limitations.. What’s worse is that this entire setup would be easier and more stable if we rebuilt it using LiveKit, which says a lot about how unnecessarily complicated and fragile your current implementation has become. Now to the most unacceptable part: throttling paid customers. We have already paid for concurrency, and throttling accounts that have paid for dedicated capacity is completely unacceptable. This defeats the entire purpose of purchasing concurrency in the first place. You need to: Immediately disable throttling for accounts that have already paid for concurrency. Clearly define and lock the behavior of your system instead of silently changing it. Stop pushing breaking changes without communication or documentation. We are not asking for favors — we are asking for the service we have already paid for. If this continues, it forces us to seriously reconsider continuing on this platform, because right now the instability and unpredictability are costing us time, money, and credibility. I expect a clear explanation and confirmation that throttling for paid concurrency accounts is disabled — immediately. https://cdn.discordapp.com/attachments/1464709859944693770/1464740052558155806/image.png?ex=697690fa&is=69753f7a&hm=e9ca2eb77cc2db57bf9bdab5bd9c38fe7c88701e8cfcb1bfd827b18ec31148b2&

@Manoj kohli I have forwarded your issue to the technical team

@Shehroze -Support @Retell You said: “If there will be things to compensate, we will.” Fine. Then be a man of your word. Let’s see how long you take to actually compensate and, more importantly, resolve the issue you created. Now, developer to developer — answer this honestly: With this level of tight throttling, how is anyone supposed to manage a concurrency of 250?????? It is illogical, technically unsound, and unethical. You sell concurrency, we pay for concurrency, and then you silently throttle usage to the point where that concurrency is unusable. That is not a limitation — that is changing the rules after taking money. And this raises a serious question: How is your product manager even doing their job? Basic product discipline says: documentation is updated first, customers are informed, breaking changes are communicated, and only then is a change rolled out.

@Shehroze -Support @Retell update please ?

You are hitting the call error limit, which is 60 errors per minute. so your create-phone-call is have some error Normal rate limit is 1000 request per 10 seconds

With the 60 calls per minutes how can we reach 150 or 250 calls same time? i want a urgent meeting on this we losing money you guys too we pay 10-12k USD per month

Hi @Prio , It looks like there may be an issue with your API body. You’re likely encountering 400-series responses or another error condition. If those errors were not occurring, you should be able to reach the expected rate limit of up to 1,000 requests per 10 seconds. I’d recommend logging the full API requests and responses to identify why errors are being returned. That should help pinpoint the root cause.