AI agent or OpenAI node keep loading data on successful workflow run

Questions
, ,
1

Describe the problem/error/question

I am building a workflow that needs ai agent or open Ai node to process some data and output a new data for me, the agent node kept kept loading data even though the workflow show succeslease and same happena to open ai node. Even when I checked the execution I have made few hours ago, it's showing same loading... Is this caused by the new updates of n8n? I went ahead to check my open ai log and I see the request there, and the output that is supposedly to be shown on my end on n8n 
(Ielect the nodes on your canvas and use the keyboard shortcuts CMD+C/CTRL+C and CMD+V/CTRL+V to copy and paste the workflow.)

Share the output returned by the last node

There output from the previous node is a 5min video transcript.

Infomation 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:
3

Welcome to n8n, @owoeye_samuel! This is a pretty specific symptom that’s worth unpacking.

When you say the node shows “loading” even after the workflow shows success, and you can see the response in your OpenAI logs, the most likely culprits are:

1. Payload size / timeout in the canvas UI: A 5-minute video transcript can be quite large as input to an AI agent, especially if it includes timestamps. The n8n canvas sometimes gets stuck rendering the output panel when the response payload is very large. The workflow actually completed successfully, but the front-end times out loading the preview. Try checking the Executions log on the left sidebar instead of the live run panel, and click into the finished execution there. You should see the full output.

2. Max token / context window issue: If the transcript is very long, the model might be hitting its context window limit. The request goes through, but the response gets truncated or the streaming doesn’t close cleanly. Check your OpenAI log for the finish_reason field, if it says length instead of stop, that’s the issue. Fix: chunk the transcript into smaller pieces before sending, or use a model with a larger context window.

3. n8n version / streaming mode: Some earlier versions had a bug where streaming responses from OpenAI would show as “loading” in the UI but actually completed. Check your n8n version and if you’re on an older one, try disabling streaming in the OpenAI node settings.

Can you share your n8n version and roughly how long the transcript text is? That would help narrow it down.