We use OpenAI, Gemini (Google), and Claude (Anthropic) models in our platform. [1] We recently rolled out GPT-5 as a model option across Clay, which is now the engine behind our formula generator and metaprompter. [2]
If you're experiencing OpenAI API integration issues, they may be due to an incorrect API key or low limits on your account. Recent updates to the Claude Sonnet 3.7 integration have doubled the estimated token usage per Claygent run, which introduces a more conservative approach but minimizes failures and enhances stability.
What specific issues are you experiencing, and can you click into the specific cell where you're seeing any error to expose more details about the error message?
Claude issues in Clay can stem from several factors. Recent updates to Claude Sonnet 3.7 integration have doubled the estimated token usage per Claygent run, which introduces a more conservative approach but minimizes failures and enhances stability.
Rate limits are particularly common with Claude and can cause runs to fail or queue indefinitely. Clay implements rate limits to ensure system stability, and these are especially noticeable for columns with heavier prompts.
If you're seeing "Run Condition not met" messages, this indicates that predefined criteria for executing the action weren't satisfied. You can click the "explain" button to see why the run condition wasn't met.
For persistent error messages, try clearing your browser cache and refreshing the view.
Can you click into the specific cell where you're seeing the Claude error to expose more details about the error message?
If you still need help resolving the Claude issues, I’m here to assist you further. Would you like to provide more details about the errors or what you’re trying to achieve so we can continue troubleshooting?
