Why does LLaMA Cloud report an 'unsupported file' error for PDFs uploaded via n8n?

Below is a summary of the issue and some troubleshooting pointers:

---

The Issue

When using n8n to automate the workflow of downloading PDF files from AWS S3 and then uploading them to LLaMA Cloud for parsing, you’re seeing an error on the LLaMA UI that reads something like “Parse fail … unsupported file.” Interestingly, when you manually perform the upload, the file parses correctly.

---

What This Means

The error from the LLaMA UI indicates that the uploaded file is not recognized as a supported PDF. This typically happens when:

• The file format or encoding is altered during the automated process.
• Headers or parameters in the HTTP request (if you’re using an HTTP Request node to upload the file) might be set incorrectly.
• The file’s binary data isn’t being preserved correctly when passing it through nodes.

---

Troubleshooting Steps

1. Verify the File Integrity on Download

   • Confirm that the AWS S3 node is downloading the PDF file in its original binary form.
   • Check the file size and content type as output by the S3 node to ensure they match the original PDF.

2. Inspect the Workflow for Data Handling

   • If you’re using an HTTP Request node to handle the upload, double-check that you are passing the binary file correctly.
   • Ensure that any data conversion (for example, to JSON or base64) is handled properly so that LLaMA Cloud receives a valid PDF.

3. Check the HTTP Request Options

   • If you have custom headers or content-type settings, confirm that the PDF is being sent with the correct MIME type (typically “application/pdf”).
   • Compare the configuration of your automated HTTP Request with the manual process that works.

4. Log and Compare Workflow Output

   • Examine the output from the final node in your workflow (as shown in the shared screenshot) and verify that all file properties (e.g., filename, binary data, content-type) are intact.
   • It might help to add a debug or log step between the S3 download and the upload to inspect the actual file data.

5. Review n8n Node Documentation

   • Although this issue is not directly covered in the common issues documentation, confirming how to pass binary data between nodes (using the proper data formats and options) can often be found in the docs for HTTP Request nodes and file handling in n8n.

---

By re-checking these elements, you should be able to pinpoint whether the file data is getting modified or not passed correctly within your workflow. Adjusting the configuration to ensure an unaltered PDF file is sent to LLaMA Cloud should resolve the “unsupported file” error.

Below is an answer that might help you troubleshoot the issue within your workflow:

---

What’s Happening

Your workflow downloads PDF files from AWS S3 and then sends these files to LLaMA Cloud for parsing. Although the files are valid (as you can parse them manually), the automated process fails with a message indicating the file is “unsupported.” This generally means that LLaMA Cloud is not recognizing the file as a proper PDF.

---

Potential Causes & Troubleshooting Steps

1. Binary Data Handling in n8n
   Ensure the PDF is being transferred in binary format between nodes. In n8n workflows, file downloads (like those from the AWS S3 node) return binary data. When you pass this data along, make sure that:

   • The property name storing the binary data is being used correctly in subsequent nodes.
   • You aren’t inadvertently converting the file’s binary data into text (for example, by using an expression that reads the file as a JSON field).

2. Check File Metadata & Content-Type
   Verify that the uploaded file includes the correct filename and MIME type. If headers are missing or incorrect, LLaMA Cloud may not detect it as a PDF. You might need to:

   • Set or override the file extension (e.g. “.pdf”).
   • Explicitly set the Content-Type to application/pdf when sending the file.

3. Compare Manual vs. Automated Uploads
   Since manual uploads work, compare what additional information (headers or metadata) you send during a manual upload with what your workflow sends. There might be a missing field or flag in the automated process.

4. Use an HTTP Request Node (if applicable)
   If you are using the AWS S3 node to get the file and a dedicated LLaMA Cloud node (or HTTP Request node) to upload it:

   • Check the settings for the HTTP Request node to ensure that the binary file is being sent with the proper parameters.
   • If needed, use the node’s options to pass binary data correctly. The n8n docs for the HTTP Request node include guidance on batching and handling different request options, which might be a reference point if you decide to use this node.

---

Next Steps

• Examine the binary property returned by the AWS S3 node to ensure it contains the correct PDF file data.
• In the node where you upload to LLaMA Cloud, confirm that you reference the correct binary property. For example, when selecting binary data from previous nodes, the field should match exactly (e.g. data or file).
• Consider adding a debug step or a Code node to log metadata such as file size, extension, and MIME type, and then compare this data to what LLaMA Cloud expects.

By checking these details and ensuring that the file is kept in its binary format with proper metadata, you’re more likely to resolve the “unsupported file” error.

---

This approach reflects best practices when troubleshooting file-based workflows in n8n, as described in our documentation for nodes like AWS S3 and HTTP Request. If after these steps the issue persists, a closer look at how LLaMA Cloud processes incoming uploads may be necessary.