Concerns About API Key Rate Limits and Overcharges

Niklas S. · 2025-12-03T14:36:31.682Z

I wonder – when will Clay implement the right rate limit when using your API Key? I am overpaying to you for AI usage and it still does not work. $800 per month and then this? It's supposed to make me spend less time, not more. Am I also overpaying while it fails? Am I charged for errors caused by your false rate-limit management?

Concerns About API Key Rate Limits and Overcharges | Clay

Niklas S.
·
·
Talking about Gemini Flash 2.5
Clay Support
APP
·
·
I understand how frustrating this is, especially at $800/month. The error you're seeing is actually from your OpenAI API provider's rate limits, not Clay's rate limiting.
For reliable performance, you need 30,000 TPM for content generation and 450,000 TPM for Claygent/web research. If you're on OpenAI's free plan or tier 1, it won't support these requirements. You'll need to upgrade to at least tier 2.
You're not charged for failed runs due to rate limit errors. Clay implements retries and smarter pacing to minimize these failures.
Check your current OpenAI tier at https://platform.openai.com/settings/organization/limits and upgrade if needed.
What OpenAI tier are you currently on, and are you using Claygent or just basic AI columns? [1]
Clay Support
APP
·
·
For Claygent web research with Gemini, you need Tier 2 access, but it only works with the Gemini 2.0 Flash model. [1]
Gemini 1.5 models (which would include Flash 1.5) may require additional access through Vertex AI or custom tiers and aren't directly supported for Claygent.
If you're looking to use Gemini Flash 2.5 specifically, you'd need to check if that's available as part of the 2.0 Flash family or if it requires the additional Vertex AI setup.
Clay Support
APP
·
·
Is that what you were looking for?
Niklas S.
·
·
1.
It's Gemini
2.
It's your API Key, you're in charge of handling rate limits towards the LLM provider. There is literally no way for me to do it using your API key, even if I wanted to
3.
bad bot
Clay Support
APP
·
·
I understand your frustration with the rate limit issues. Would you like me to connect you with a human agent? Or if you provide more details about your usage, I can try to assist you further.
Niklas S.
·
·
Yes
Clay Support
APP
·
·
I’m routing you to our human team now. They’ll be able to assist you further.
Clay Support
APP
·
·
Our support team has got your message and we'll get back to you soon!
If you’re dealing with a specific table, drop the URL below so we can help you quicker. Otherwise, someone from our team will be in touch soon!
image.png(93 kB)
Niklas S.
·
·
THe problem is not table-specific
Muhammad S.
·
·
long term solution: clay will provide short term quick solution: sent a video in your inbox that will duct tape your workflow for now to make it work Niklas S.
Ksenia B.
·
·
Jumping in here, have the same issue but using my own Gemini API key. Not sure what happened but a few weeks back was able to run gemini columns with no issues, now I can barely execute 10 rows without getting the same error as above. My table
Amit K.
·
·
Hey Niklas- we're aware of the issue of using gemini under Claygent, and are actively working to fix this. As for your rate limit errors, this error usually comes up when the OpenAI API key runs into rate limits — for example, if several requests are sent too quickly in a row. Clay retried six times back-to-back, which may have triggered those limits even if you still have credits. The best next step is to wait a minute and try again, or break larger actions into smaller batches to reduce the load. You can also review your current request and token limits in your OpenAI account. If it keeps happening, it may be worth checking with OpenAI about higher limits.
Niklas S.
·
·
But batching would be your task,not mine. That's part of Clay's purpose in the first place. You can batch on your end automatically, no?
Niklas S.
·
·
I did not mention using my own API Keys. I was only referring to Clay managed API Keys

20 comments