Manual install TEXT GEN WEBUI

Manual Install Method from my last Youtube video: https://youtu.be/QSJS9YDqf9w


🧠 HOW TO INSTALL TEXT-GEN WEBUI MANUALLY (STEP-BY-STEP)

───────────────────────────────
💡 WHAT THIS DOES
───────────────────────────────
This guide explains how to:
1. Install Text Generation WebUI (Oobabooga)
2. Launch it locally or on RunPod
3. Download a GGUF model manually from Hugging Face

No Patreon scripts required — just follow along.


───────────────────────────────
⚙️ 1. INSTALLING ON WINDOWS
───────────────────────────────

STEP 1. Install Prerequisites
-----------------------------
Make sure you have:
- Git → https://git-scm.com/download/win
- Python 3.10 or 3.11 (64-bit) → https://www.python.org/downloads/windows/
During Python install, check the box “Add Python to PATH.”


STEP 2. Clone the WebUI
-----------------------
Open Command Prompt (CMD) and run:

    git clone https://github.com/oobabooga/text-generation-webui.git
    cd text-generation-webui


STEP 3. Start the Installer
---------------------------
Run:

    start_windows.bat

This will:
- Create a Python virtual environment
- Install dependencies
- Launch the WebUI

When you see:

    Running on local URL: http://127.0.0.1:7860

open that address in your browser.


STEP 4. First Launch
--------------------
On first startup, the WebUI will ask what loader to use.
Type:

    A

and press Enter (this selects “llama.cpp”).


───────────────────────────────
☁️ 2. INSTALLING ON RUNPOD (LINUX CLOUD)
───────────────────────────────

STEP 1. Open a Terminal
If you’re inside a RunPod container, open the terminal.

STEP 2. Clone the Repository

    git clone https://github.com/oobabooga/text-generation-webui.git
    cd text-generation-webui


STEP 3. (Optional) Add Login Protection
Run this to protect your WebUI:

    read -p "Username: " user; read -p "Password: " pass
    export AUTH="--gradio-auth ${user}:${pass}"


STEP 4. Launch the WebUI
------------------------
Run:

    printf '%s\n' "A" | ./start_linux.sh --listen --listen-host 0.0.0.0 --listen-port 7860 $AUTH

Then open the link shown (for example https://XXXXXX.runpod.run:7860).


───────────────────────────────
🧩 3. DOWNLOAD A MODEL MANUALLY (HUGGING FACE)
───────────────────────────────

STEP 1. Create a Folder
-----------------------
Inside your WebUI folder, create this path:

    text-generation-webui/user_data/models


STEP 2. Go to Hugging Face
--------------------------
Find a model you like, for example:
https://huggingface.co/unsloth/GLM-4.5-Air-GGUF/tree/main/Q8_0


STEP 3. Download the .GGUF File
-------------------------------
Click the .gguf file(s) and press “Download”.
Place the file(s) inside:

    text-generation-webui/user_data/models/


STEP 4. Load It in WebUI
------------------------
1. Go to the “Model” tab
2. Choose Loader: llama.cpp
3. Select your .gguf file
4. Click “Load model”

Done! You can now chat locally with your LLM.


───────────────────────────────
⚠️ COMMON ISSUES
───────────────────────────────

Q: Torch not found / missing module?
A: Delete the folder “installer_files/env” and rerun start_windows.bat.

Q: The WebUI closes instantly?
A: Open CMD manually, cd to the WebUI folder, then run:
       start_windows.bat
to see the full error.

Q: Multiple .gguf shards?
A: Keep all parts (e.g. -00001-of-00003.gguf) in the same folder.


───────────────────────────────
✅ TL;DR QUICK COMMANDS
───────────────────────────────

Windows:
---------
    git clone https://github.com/oobabooga/text-generation-webui.git
    cd text-generation-webui
    start_windows.bat

Linux / RunPod:
---------------
    git clone https://github.com/oobabooga/text-generation-webui.git
    cd text-generation-webui
    printf '%s\n' "A" | ./start_linux.sh --listen --listen-host 0.0.0.0 --listen-port 7860

Then:
- Put your .gguf files into user_data/models
- Load them in WebUI → Model tab → Load


───────────────────────────────
❤️ OPTIONAL AUTH (LINUX)
───────────────────────────────
    printf '%s\n' "A" | ./start_linux.sh --listen --listen-host 0.0.0.0 --listen-port 7860 --gradio-auth user:pass