@GollyJer - my hunch here is that this simply doesn’t work. I tried everything I could, but couldn’t get it to work.
My only option was to invoke lambda functions from my n8n workflow (where the authentication can be configured as I need) and then interact with the relevant service e.g. Textract from within the function.
Turns out it pushed me to find a better solution.
I’m now using Gemini 2.0 Flash for an initial scan & Gemini 2.0 Pro + OCR.space for a validation scan and getting much better results than we ever achieved previously.
Makes me feel better to see someone else facing the same issue. I’ve spent a lot of time trying to get Textract to work in an http node with no luck. I’m also using OCR space, but it has relatively low success/accuracy rate for my use case so I will try Gemini 2.0 Flash. Thanks!
Hi @Jorge_Castellote. My use case is pretty specific but here’s a screenshot of the flow. I think most people can get away with just the first pass to the LLM Scan - Images Only.
The passes through OCR.space and the second LLM, for us, were because Gemini 2.0 Flash doesn’t do the best with small blurry text and OCR.space could sometimes pick up the words that Gemini couldn’t. From the red box and after is probably not necessary for most use cases.
Would just like to add, I’m having the same problems as well with the HTTP node and using my AWS Credentials. Same as OP, I’m also trying to use Transcribe.
I know my credentials are correct as when I use any other AWS node (such as S3, which I have earlier in my flow) it works correctly.
Hey all I found creating a Lambda function as a proxy then using the Lambda node to call it works. the AWS Signature Version 4 authentication does not work in the http request node.
