.png){kind=logo}
Issues with PDF URLs in Claygent and OpenAI/Anthropic
Is anyone else facing errors when reading out PDF URLs with Claygent and OpenAI/Anthropic? Neon etc. still works fine https://app.clay.com/shared-table/share_6KWHV6agA6oR
If you're experiencing issues with OpenAI/Anthropic integrations, this could be related to API plan limitations. For reliable use of OpenAI in Clay, you need 30,000 TPM for ChatGPT columns, and 450,000 TPM for Claygent functionality.To resolve this: 1. Check your OpenAI API key at platform.openai.com/account/api-keys 2. Verify your API usage tier and limits 3. Consider upgrading to a higher tier for better performanceIt's worth noting that Claygent Neon is our most capable model and has shown significant improvements in parsing PDFs and handling complex web research tasks with higher accuracy compared to other model options. Source: Introducing Claygent Neon
Did that answer your question?
Hi Nicolas S.! This thread was recently closed by our Support team. If you have a moment, please share your feedback:
we never had problems with this before - we have been using claygent for months and spend more than 600$ each
There have been recent changes to how AI functionality works in Clay. In October 2024, Clay consolidated all AI experiences across Claygent, Anthropic, OpenAI, and Gemini into a single "Use AI" enrichment.For the PDF URL reading functionality specifically, you'll need to ensure: 1. Your OpenAI API key is correctly configured and active 2. Your plan meets the required usage limits - specifically 450,000 TPM (tokens per minute) for Claygent functionality 3. Your API usage tier is sufficient for your current usageIf you're seeing errors, first verify your API key at platform.openai.com/account/api-keys and check your current usage limits. Even if you've spent $600, you may need to check if your OpenAI account's usage tier meets the requirements for the consolidated AI features. Source: AI action consolidation
Is that what you were looking for?