N8N HTTP Request Node gibt HTTP 500 "LLM call failed" bei Self-hosted LLM API

Questions
,
1

Setup

  • External API: alina-controller.fly.dev/api/generate (Custom LLM proxy)
  • Config: POST, Bearer Token, JSON Body, Timeout 60s, Retry 3x
  • Node: HTTP Request v4.2, N8N Cloud 1.91.2

Following works: :white_check_mark:

  • PowerShell: API calls successful with same parameters
  • Services: Both Controller and Ollama services running
  • Authorization: Bearer token correct (verified via PowerShell)
  • Request Format: JSON body validated

Following does not work :x:

  • N8N Node: Consistent HTTP 500 errors
  • Error Message: "LLM call failed", "Request failed with status code 500"

Debug Steps Tried

  1. Timeout/Retry: Increased to 60s, retry 3x
  2. Service Warm-up: Manual service startup via PowerShell
  3. Parameter Verification: All parameters match working PowerShell request
  4. Simple Test: Hardcoded prompt instead of dynamic data

Current Status

PowerShell API calls work perfectly with identical parameters, N8N fails consistently. Seeking guidance on N8N-specific request handling differences.

Request Code that works in PowerShell but fails in N8N:
{
“model”: “mixtral:8x7b”,
“prompt”: “Extract tool names from: Test text”,
“stream”: false
}

Thanks for your support

TestUpload
TestUpload1600×818 167 KB

2

I feel like it would best to look at the logs from the self hosted api itself.