Anthropic API Rate Limit Error in Clay: Need Help Fixing It

Clay S. I'm using my Anthropic API in Clay without issue until today where I got an error saying the Rate Limit for the API Key for a specific column is insufficient. This is new. I'm Tier 2 for Anthropic and according to this document https://community.clay.com/x/support/spn8m8bupiej/using-your-own-anthropic-api-key-with-clay-require it doesn't matter which tier you're on. How can I get this fixed?

25 comments

Channeled
APP
·
·
We're sorry you're experiencing this issue with your Anthropic API integration. Even with a Tier 2 account, you might still encounter rate limit errors if processing large amounts of data simultaneously. To resolve this, try processing your data in smaller batches, around 50 rows at a time. This approach can help manage your API usage within the set limits. For more detailed information about your current limits and usage, you can check your Anthropic API settings at Anthropic API settings. Additionally, you might find the Anthropic API rate limits documentation helpful for understanding the specifics of your tier's limitations. If you need further assistance or if this solution doesn't resolve your issue, please request human support by clicking on "Talk to a person" in this chat.
Stefano
·
·
Hi Scott, thank you for reaching out. A rate limit means that you were trying to run too many cells in a too small period of time. This means that to get around this error you will need to wait a bit of time and then re-run the cells. Let me know if this helps.
Scott Y.
·
·
Hi Stefano for your quick reply. I turned the table off and tried running one record and am still getting the error. I've run out of Clay credits so not sure if this is the issue (even though my table doesn't use any).
Stefano
·
·
Hi Scott, I suggest waiting a bit more time before running the enrichments. A rate limit means that you were trying to run too many cells in a too small period of time. This means that to get around this error you will need to wait a bit of time and then re-run the cells. Let me know if this helps.
Scott Y.
·
·
Hi Stefano I waited 12 hours and only running 1 cell won't now work. I'm at a loss here...
Scott Y.
·
·
Unfortunately I'm also not at a high enough tier with OpenAI to use their API.
Daniel K.
·
·
Hey Scott, thanks for reaching out! It looks like you’re running into an issue with rate limits. Due to the low limits on Anthropic keys, Claygent currently runs reliably only on Tier 4 (400k TPM). We’re exploring support for Tier 3, but for now, Tier 4 is required. As a workaround, you could use Claycredits with Anthropic, switch to a Gemini API key, or consider upgrading your OpenAI tier. Sorry for the hassle—let me know if you have any follow-up questions!
Scott Y.
·
·
Ok. Which Perplexity and Gemini Tiers do you need for them to work?
Bo (.
·
·
Hey Scott, To run Gemini inside Claygent, you’ll typically need a tier that supports around 400k tokens per minute. Perplexity, unfortunately, isn’t supported in Claygent at the moment. Let me know if you want help finding the right Gemini tier.
Scott Y.
·
·
Bo (. Seems like Gemini is working, thanks on Perplexity, and OpenAI doesn't seem to be working even at Tier 2.
Owen C.
·
·
Hey Scott, thanks for reaching out. Wanted to check in here, is Open AI still giving you issues even though you're on a tier 2 plan? If so, do you mind sending the link (url) to the table so we can take a look?
(52 kB)
Michael S.
·
·
Hey I am struggling with a similarrrr issueeee
Michael S.
·
·
just trying to figure out how to get to the tier 2 plan in the first place
Bo (.
·
·
Hey, thanks for reaching out! 😊 I’m really sorry you’re running into this issue. The error message you’re seeing is related to trying to connect to your OpenAI key, and the reason is that your key needs to be upgraded to a higher tier to function properly. Unfortunately, we can’t bypass that limit—it’s something OpenAI handles directly. Here’s what you need for using your API key with the integrations: Claude You need to be on Tier 4 or higher to run Claygent with Claude. Because that's the only tier able to support the required token per minutes. OpenAI You’ll need at least 450,000 Tokens Per Minute (TPM) to run the Claygent Text Generation If you’re just generating text with existing data in Clay, you only need 30,000 TPM. That said, Clay credits will work for these tasks, so upgrading your OpenAI tier might not be necessary depending on your needs. You can check your current tier, as well as your usage and limits, in the OpenAI API dashboard here: • Usage • Limits • Usage Tiers Or Claude Console: https://console.anthropic.com/ We’ve seen users successfully add a pre-paid tier amount to their OpenAI account and resolve the issue. However, since this is outside our control, we recommend troubleshooting directly with OpenAI’s team. Let me know if you need any further assistance! 😊
Scott Y.
·
·
Michael S. I had to actually spend $50 in OpenAI's API to get moved to tier 2. I created a Make automation to burn through the credits. Basically had to waste about $10 to do it ...
Michael S.
·
·
haha yeah I just wrote my own automation in python to get to the spend limit and now i'm tier 3, thank you!
Scott Y.
·
·
Awesome and congrats! Sounds like you're an overachiever lol.
Michael S.
·
·
haha trying
Michael S.
·
·
now I'm at Tier 4 time to make clay go brrrr
Michael S.
·
·
sounds like we're starting at a similar time, let me know if you ever want to show and tell on what you're working on!
Scott Y.
·
·
Michael S. Looks like I'm 2 hours behind ya. I'm using Clay for a cold email automation following Lead Gen Jay's stuff. You?
Michael S.
·
·
DMing you!
Channeled
APP
·
·
Dane V.
·
·
Has anyone on Tier 4 managed to resolve the rate limit issue when using Claude 3.7 Sonnet? I am on tier 4 and last week it was working fine but now I am getting the rate limit issue when using 3.7 Sonnet. Is there a reason for this? The Anthropic API doc says the ITPM for 3.7 Sonnet is only 200,000. Based on Daniel K. response, they only support 400k at the moment. When I use 3.5 Sonnet it works but I prefer the results I was getting with 3.7 Sonnet. Any help to get Claude 3.7 Sonnet working again would be great.
Bo (.
·
·
Hey — sorry about the trouble. Just to give you some context: we had to make a change after seeing many users burn through credits by using Claude with a lower model tier, which ended up costing them significantly. Claude 3.7 Sonnet now requires 400,000 tokens per minute, and unfortunately, there’s no way for us to bypass that limit. Very sorry for this