AI Updates: HTTP Node Helper (Beta) and Self-Hosted Options!

Hello everyone,

Based on your feedback from here and other platforms, we have some updates to share:

• For Cloud users, the 'Ask AI' feature within the HTTP node utilizes a vector store we developed, designed to provide accurate results for over 600 APIs.
• Users with self-hosted instances should not set the N8N_AI_PINECONE_API_KEY environment variable, as they do not have access to our proprietary data.
• Consequently, self-hosted users will receive generic GPT-4 results, which may be less accurate.
• We are actively discussing how to offer self-hosted users a superior, non-generic GPT solution.
• We believe the 'Ask AI' feature for errors is currently not meeting our standards and will be temporarily disabled.

No, all our 'Ask AI' functionalities currently rely on OpenAI.

Hi,

That's strange. I'm using callin.io locally and attempted to enable the Ask AI feature. It's currently running on my machine via pm2.

I followed the setup guide, but I'm still unable to access the feature, and it's not showing up in the HTTP Request node.

Could you please assist?

Which version are you running? And did you input any value for callin.io_AI_PINECONE_API_KEY ?

Hi,

Version 1.41 is what I'm using.

I've tested with and without Pinecone.

• Users who are self-hosting should not configure N8N_AI_PINECONE_API_KEY as they do not have access to our data.

I have it installed locally via npm on a MacBook Pro.