Convrsational agent on open blog posts resulting exceed token error

[Kunal_Malhotra]

Error - Request too large for gpt-4o in organization org-e4yW1j9yPVVoLVlKMvWB0s6b on tokens per min (TPM): Limit 30000, Requested 36289. The input or output tokens must be reduced in order to run successfully. Visit https://platform.openai.com/account/rate-limits to learn more.

i have built a conversational agent by parsing a few blog posts. I am using Supabase Vector Store and OpenAI Embeddings. Though I am sending all the embeddings to the Agent, which could be the reason for the error above.

But how to build an agent that looks at 1000s of embeddings in the vector database ?

• n8n version: 1.69
• Database (default: SQLite): Supabase postgres
• n8n EXECUTIONS_PROCESS setting (default: own, main): default
• Running n8n via (Docker, npm, n8n cloud, desktop app): n8n cloud
• Operating system:

[n8n]

It looks like your topic is missing some important information. Could you provide the following if applicable.

• n8n version:
• Database (default: SQLite):
• n8n EXECUTIONS_PROCESS setting (default: own, main):
• Running n8n via (Docker, npm, n8n cloud, desktop app):
• Operating system:

[Kunal_Malhotra]

@Jim_Le been following all the great work you do. Would appreciate your guidance on how to build such a solution

[Jim_Le]

Hey @Kunal_Malhotra

Hmm the point of the vector store is that you don’t have to send so many documents to the LLM. You want the vector store to pick out the most relevant to your user’s query so top 4-5 and not 1000.

If your goal is to recommend a huge number of blog post urls back to the user, you don’t necessarily need the LLM. Here’s an example where as long as your last node has an “output” json, it’ll reply to the chat.