LLM Timeout Diagnostic Suite

Purpose: Test LLM server (192.168.3.40:8080) with SDXL HD cold start (~65s) across different tech stacks

Instructions: Click each test button ONE AT A TIME. Wait for completion before running next.

New: Each test streams heartbeat bytes every 20s back to the browser. The raw chunk log shows exactly what arrives and when.

PHP Test

Raw chunk log (shows heartbeats + final response):

        
Parsed result:

        
Not yet run

    


    

        
Python Test

        
        
        
Raw chunk log (shows heartbeats + final response):

        

        
Parsed result:

        
Not yet run

    


    

        
Node.js Test

        
        
        
Raw chunk log (shows heartbeats + final response):

        

        
Parsed result:

        
Not yet run

    


    

        
Browser JavaScript Test (Direct to LLM Server)

        
        
        
Raw chunk log (heartbeats come directly from LLM router):

        

        
Parsed result:

        
Not yet run