“must contain a binary file ‘data’, but none was found [item 0]”

Questions
1

Trying to paste txt from xlsx and two pdfs to Anthropic AI to work, but stuck at “must contain a binary file ‘data’, but none was found [item 0]”

“must contain a binary file ‘data’, but none was found [item 0]”

Please share your workflow

Share the output returned by the last node

Got 4 of 4 output at the last code merge, all returning what I want to pass onto AI model API. However at the final stage it always come back with “must contain a binary file ‘data’, but none was found [item 0]” or PDF cannot be found by AI’s return.

Docker environment with Windows11.

Information on your n8n setup

  • n8n version:
  • Database (default: SQLite):
  • n8n EXECUTIONS_PROCESS setting (default: own, main):
  • Running n8n via (Docker, npm, n8n cloud, desktop app):
  • Operating system:
2

Please run the workflow, open the “Message the Model” node and show us the input.

4

Can you try this version? I have:

  • added an extra merge node - it forces all previous branches to complete before continuing to the AI node. In your flow the binaries are not ready but the text branch activates the AI node

  • split the binaries into multiple items after your final merge(just for testing)

  • changed how you pass them in the AI node so you can pass an array with all of them without using names

Download the JSON here, as it’s too long: https://drive.google.com/file/d/1hDqA0P7qBoTTklCVek8ekgzfQvS2Y62A/view?usp=sharing

Also, if you wish to optimize the huge number of Extract nodes, you can check this solution:

1 Like
5

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.