⚡ Single-File PHP AI Chat (Ollama & OpenAI)

A full-featured AI chat interface that lives in one single PHP file.

No npm install. No docker compose. No build steps. Just drop chat.php on any server (or localhost) and start chatting with Llama 3 (via Ollama) or GPT-4 (via OpenAI).

🚀 Why this exists?

Modern AI tools are bloated. Running a simple UI for your local LLM often requires installing Node.js, Python, React, and 2GB of dependencies.

This project solves that:

• Zero Dependencies: Uses vanilla PHP, jQuery (CDN), and CSS.
• Portable: Runs on any shared hosting (Apache/Nginx) or a local potato.
• Private: Your data stays on your server (or localhost).
• Streaming: Real-time token streaming via Server-Sent Events (SSE).

✨ Features

• Universal Support: Works with Ollama (Local) and OpenAI (Cloud).
• Code Highlighting: Auto-detects code blocks and applies syntax highlighting.
• Markdown Support: Renders rich text, tables, and lists.
• Chat History: Saves your conversations locally (in browser localStorage).
• Secure: Simple password protection included.
• Responsive: Works perfectly on mobile and desktop.

🛠️ Quick Start

Option A: Local Llama (Ollama)

1. Ensure Ollama is running.
2. Download chat.php.
3. Open the file and edit the configuration at the top:

   $apiKey  = 'ollama'; 
   $baseUrl = '[http://127.0.0.1:11434/v1](http://127.0.0.1:11434/v1)'; // Use 127.0.0.1 for local reliability
   $model   = 'llama3:8b'; // Match the model you pulled

4. Run a local PHP server in your terminal:

   php -S localhost:8000

5. Open http://localhost:8000/chat.php

Option B: OpenAI (ChatGPT)

1. Download chat.php.
2. Edit the configuration at the top:

   $apiKey  = 'sk-proj-...'; // Your OpenAI Key
   $baseUrl = '[https://api.openai.com/v1](https://api.openai.com/v1)';
   $model   = 'gpt-4o-mini';

3. Upload to any PHP web host.

⚙️ Configuration

All settings are located at the very top of chat.php:

// Security
define('ACCESS_PASSWORD', 'secret'); // ?pwd=secret to login

// API Config
$apiKey  = '...';
$baseUrl = '[http://127.0.0.1:11434/v1](http://127.0.0.1:11434/v1)'; 
$model   = 'llama3';

☕ Support the Project

If you found this tool useful, you can buy me a coffee to keep me awake while I code the next feature!