Custom AI Chat Multi modal interface made in javascript/php, runs on apache or nginx, ie. for windows use xamp with a few pre configurations for long running tasks-

*You need to supply your own AI backend to the interface, ie. Ollama+Stable Diffusion for reason+vision+diffusion and set BACKEND_IP in setup.php

—
Pre-configurations for xamp - once installed, open XAMP control panel, only ever need Apache (no mysql etc etc), open config -> apache2.conf and add to the bottom of the file:

Timeout 1800

—
Then open config again -> php.ini and change these variables to these values:

max_execution_time=1800
memory_limit = 512M
post_max_size=5050M
upload_max_filesize=5000M
max_file_uploads=50

extension=gd ;uncomment in conf to enable for thumbnail/preview creations

—
Then Stop/Start and request to AI should no longer time out/invalid json if cut short if it takes 10-20 min. (If the AI backend is running on a slow/old system, may need to increase 1800 to 3600 etc, especially for vision model!)

(For nginx see details in note.txt about fcgi timeouts supplied with zip)

—

Use the setup_and_launch.py on the backend computer to setup/run the AI models (script is lacking the download parts for setup in both ollama and stable diffusion but can install through command line like this, or give the script to ai to complete to your needs:

Linux:
curl -fsSL https://ollama.com/install.sh | sh

Windows:
winget install -e --id Ollama.Ollama

To upgrade later
winget upgrade -e --id Ollama.Ollama