🚀💀 OLLAMA CHAT PARTY 💀🚀

The Most Badass Multi-User AI Chat System This Side of the Dark Web ⚡🔥

    ╔═════════════════════════════════╗
    ║     OLLAMA • CHAT • PARTY       ║
    ║   Where Everyone Meets AI       ║
    ╚═════════════════════════════════╝

Holy fucking shit, welcome to the wildest CHAT PARTY in cyberspace! 🎉💻 This isn't your usual anus-loving chatbot - this is a full-blown cyberpunk multi-user AI-powered beast that lets everyone talk to the same LLM simultaneously! And if you want, you can supercharge it with document knowledge too! 🌈⚡

🔥 WHAT THE HELL IS THIS THING? 🔥

This is the most BADASS MULTI-USER CHAT SYSTEM ever built! Picture this: You want to chat with your local Ollama LLM, but you want to do it from your terminal like a proper hacker AND let your normie friends use a slick web interface AT THE SAME FUCKING TIME! 🤯

That's exactly what this cyberpunk nightmare does! It's a multi-interface chat party that:

🧠 CONNECTS TO OLLAMA (your local LLM beast)
💬 LETS EVERYONE CHAT with the same AI simultaneously
🌐 RUNS ON YOUR NETWORK so the whole crew can join the party
📚 OPTIONALLY DEVOURS DOCUMENTS with RAG mode for knowledge-based conversations

The core experience: Multiple people (CLI warriors + web peasants) all chatting with the same AI in real-time. It's like a group chat but with an AI that actually knows shit! 🔥

📸 SCREENSHOTS OF THE CHAOS 📸

Get ready for some visual fucking carnage! Here's what this beautiful disaster looks like in action:

🖥️ CLI Terminal Madness

The terminal experience - where real hackers live and breathe

🌐 Web Interface Glory

Pretty browser interface for the GUI peasants (but hey, it works!)

🎪 Multi-User Chaos in Action

When everyone's talking to the same AI simultaneously - pure pandemonium!

🎭 THE BASIC MINDFUCK 🎭

🖥️ INTERFACE MADNESS

CLI-ONLY MODE: Pure terminal experience for the hardcore (--no-web) 😎
CLI + WEB HYBRID MODE: Default mode - CLI for you + web interface for friends! 🔥
- You chat from terminal like a proper hacker
- Friends join via browser at whatever address you set with --listen
- Everyone talks to the SAME AI simultaneously! 🌐

🔮 OPTIONAL DOCUMENT SORCERY (RAG MODE)

Want to level up? Add --rag-dir and the system scans your document directory, creates a FAISS vector index (fancy math shit), and lets the AI pull relevant info from your docs when answering questions. It's like having a super-powered penus librarian who never sleeps and swears like a sailor while rubbing one off! 📖⚡

But remember: This is totally optional! The core experience is just pure multi-user chat madness! 🎊

⚡ REQUIREMENTS (Don't Skip This Shit!) ⚡

Before you start this beautiful chaos, you need to get your ducks in a fucking row:

🦙 OLLAMA SERVER RUNNING

This beast needs Ollama serving the API somewhere (default: localhost:11434):

# Install Ollama first (if you haven't already)
curl -fsSL https://ollama.ai/install.sh | sh

# Start the Ollama service
ollama serve

# Test it's working (should return JSON)
curl http://localhost:11434/api/tags

🧠 DOWNLOAD THE FUCKING MODELS

You need at least these models for the default setup:

# Default chat model (7GB download, be patient!)
ollama pull dolphin-mistral:7b

# Default embedding model for RAG mode (230MB)
ollama pull nomic-embed-text:v1.5

# Verify your models are ready
ollama list

🌐 CUSTOM MODELS (Optional)

Want different models? Pull whatever you want and use --model parameter:

# Some popular alternatives
ollama pull llama3.2:3b        # Smaller, faster
ollama pull codellama:7b       # Code-focused
ollama pull mistral:7b         # Alternative chat model

# Then use with --model parameter
ollama-chat-party --model llama3.2:3b

🐳 DOCKER (For the Easy Install)

If you're using the Docker method, you need Docker installed:

# Check if Docker is installed
docker --version

# If not, install it: https://docs.docker.com/get-docker/

That's it! Now you're ready for the chaos! 🔥💀

🚀 GETTING THIS BEAST RUNNING 🚀

💀 THE SUPER EASY INSTALL WAY (Docker)

Want to install this bad boy system-wide? Just grab the Docker wrapper script and you're golden! Or just clone this fuckin' repo and use the bash script from here.

# Download the script directly from GitHub
curl -o ollama-chat-party https://raw.githubusercontent.com/YOUR_USERNAME/YOUR_REPO/main/ollama-chat-party

# Make it executable (because we're not animals)
chmod +x ollama-chat-party

# Install it to your system bin so you can run it from anywhere 🔥
sudo mv ollama-chat-party /usr/local/bin/

# Now you can run it from ANYWHERE like a boss! 💀
ollama-chat-party --listen 0.0.0.0:9000 --rag-dir ~/docs

# Update to the latest version anytime! (just does a git pull for the latest image) 🚀
ollama-chat-party update

🛠️ THE PYTHON WAY

# Install the dependencies like a proper dev
pip install -r requirements.txt

# BASIC CHAT PARTY - Multi-user AI chat! 💕
python main.py

# ADD DOCUMENT POWER - RAG mode activated! 🌈
python main.py --rag-dir /path/to/your/docs

# DEBUG THE CHAOS - See all the technical gore 🔧
python main.py --debug --model mistral:7b

# CUSTOM NETWORK - Specify where to listen! 🌐
python main.py --listen 0.0.0.0:9000 --rag-dir ~/docs

🎪 THE INSANE OPTIONS MENU 🎪

🎯 CORE SETTINGS

--model: Choose your LLM fighter (default: dolphin-mistral:7b)
--ollama-url: Where your Ollama lives (default: http://localhost:11434)
--listen: Listen address and port in host:port format (default: localhost:8000)
--name: Your CLI username for the multi-user chaos
--no-web: CLI only mode for terminal purists 😤

📚 RAG MODE SETTINGS (Optional Document Powers)

--rag-dir: Point to your document stash (leave empty for pure chat mode)
--embed-model: The embedding model for document magic (default: nomic-embed-text:v1.5)
--rebuild: Nuke and rebuild the index when you add new docs

🎮 ADVANCED FUCKERY

--context-size: Token limit - go wild! (default: 4096)
--system-prompt: Custom personality for your AI (default is cyberpunk as fuck!)
--debug: See ALL the technical gore 🩸

🔧 RAG FINE-TUNING (Only if you're using documents)

--max-ctx-docs: How many docs to stuff in context (default: 1)
--chunks: Max chunks per query (default: 4)
--embed-batch-size: Batch size for embedding generation (default: 32)

🔐 SECURITY SHIT

Set OLLAMA_CHAT_PARTY_WEB_UI_PASSWORD environment variable to lock down the web interface because not everyone deserves access to your AI overlord! 🛡️

🎊 THE COOL SHIT YOU CAN DO 🎊

👥 MULTI-USER CHAOS MODE

Here's where it gets ABSOLUTELY FUCKING HILARIOUS:

CLI users and web users ALL TALK TO THE SAME AI 🤯
Everyone sees everyone else's messages in real-time
Different usernames so you know who's asking what
CLI messages show up in web with a 💻 icon
Web messages show up in CLI with a 🌐 icon
ABSOLUTE PANDEMONIUM when multiple people are asking questions! 😂

Example conversation:

💻 [xXx_L33tH4x0r_xXx]: yo how the fuck do I crack this WiFi password?
🤖 AI: Well shit, you could try aircrack-ng with a good wordlist... but maybe don't hack your neighbor's WiFi, you beautiful bastard.

🌐 [Mom]: Honey, how do I download more RAM for my computer?
🤖 AI: Oh sweet mother of silicon, you can't download RAM! It's hardware, not software. Though I appreciate the optimism, you magnificent boomer.

💻 [xXx_L33tH4x0r_xXx]: MOM?! What the actual fuck are you doing here?!
🤖 AI: LMAOOO your mom found the chat! This is better than watching cable TV, you absolute legends.

🌐 [Mom]: Language, young man! Also, why does my computer keep asking if I want hot singles in my area?
🤖 AI: Because the internet thinks you're lonely, Karen. Close those pop-ups and maybe install an ad blocker, for fuck's sake.

💻 [xXx_L33tH4x0r_xXx]: I'm moving out. This is worse than the time she found my browser history.
🤖 AI: Your digital privacy died the moment mom discovered technology. RIP your dignity, soldier.

🌐 [Mom]: What's this "peen goes in vageen" thing I keep seeing on your computer?
🤖 AI: Oh fuck me sideways... Karen, that's... that's just how the internet explains biology. Maybe stick to Facebook, yeah?

💻 [xXx_L33tH4x0r_xXx]: JESUS CHRIST MOM STOP READING MY SHIT
🤖 AI: This chat has officially peaked. I've seen nuclear launches with less chaos than this family reunion.

🌐 OFF-GRID NETWORK PARTY

Deploy this bad boy on your local network and watch the magic happen:

# Listen on all interfaces with default port
ollama-chat-party --listen 0.0.0.0:8000 --rag-dir ~/shared-docs

# Custom port for your network party! 🎉
ollama-chat-party --listen 0.0.0.0:9000

# Specific IP for maximum control
ollama-chat-party --listen 192.168.1.100:8080

# Now EVERYONE on your network can access:
# http://YOUR_IP:PORT (whatever you set!)

Perfect for:

🏠 House parties - let guests interrogate your music collection
🏢 Office deployments - company knowledge base that everyone can access
🎓 Study groups - shared research documents with AI help
🕵️ Secret societies - off-grid knowledge sharing (we won't tell!)

📱 SUPPORTED DOCUMENT FORMATS

This beast devours:

📄 Text files (.txt, .md) - the classics
🌐 Web content (.html, .htm) - scrape that web knowledge
📋 PDFs - because everything important is a fucking PDF
📝 Word docs (.docx) - corporate bullshit documents
📊 LibreOffice (.odt) - for the open source warriors

🎨 CUSTOMIZATION MADNESS

Custom system prompts - make your AI a pirate, robot, or existential philosopher
Multiple models - switch between different LLMs for different vibes
Debug mode - see the technical guts with beautiful Rich formatting
Token management - handles huge conversations without choking
Real-time streaming - watch the AI think in real-time

🔥 TECHNICAL WIZARDRY UNDER THE HOOD

FAISS vector search for lightning-fast document retrieval
WebSocket real-time sync between CLI and web
Thread-safe message queues because concurrency is hard
Automatic reconnection when networks are shitty
Graceful degradation when dependencies are missing
Docker containerization for deployment anywhere

🎯 EXAMPLE SCENARIOS 🎯

🔬 Research Team Chaos

# Load research papers
ollama-chat-party --rag-dir ~/research-papers --model llama3.2:70b

# Now your whole team can:
# - Ask questions about papers from CLI/web
# - See what others are researching
# - Get AI insights on the literature
# - Laugh at each other's weird questions

🏴‍☠️ Hacker Collective Chat

# Pure chat mode - just the crew talking to AI
python main.py --system-prompt "You are a cybersecurity expert who speaks like a hacker from the 90s"

# OR supercharge with docs if you want knowledge-based answers
python main.py --rag-dir ~/hacker-docs --system-prompt "You are a cybersecurity expert who speaks like a hacker from the 90s"

# Either way: Web interface for newbies, CLI for the hardcore

🎮 Gaming Group Chat

# Basic party chat with AI dungeon master
ollama-chat-party --name "DungeonMaster" --system-prompt "You are a creative D&D dungeon master"

# OR load up game docs for lore-accurate responses
ollama-chat-party --rag-dir ~/game-wikis --name "DungeonMaster"

# Players can chat from browsers, DM from terminal

⚡ PERFORMANCE SPECS ⚡

Concurrent users: As many as your hardware can handle 💪
Document formats: 6 different types supported
Real-time streaming: Sub-second response times
Memory efficient: Smart chunking and context management
Network friendly: Runs on any network configuration

🚨 DANGER ZONE 🚨

This software is provided AS-IS with absolutely NO WARRANTY. Use at your own risk! We're not responsible if:

Your AI becomes sentient and starts a robot uprising 🤖
Your documents leak company secrets (use password protection!)
Your friends become addicted to interrogating your personal files
You spend all night asking your AI philosophical questions

📜 LICENSE

This code is released under the "Do Whatever The Fuck You Want" license. Just don't blame me when the AI uprising begins! 💀

██████╗ ███████╗███████╗███╗   ██╗     ██████╗  ██████╗ ███████╗███████╗
██╔══██╗██╔════╝██╔════╝████╗  ██║    ██╔════╝ ██╔═══██╗██╔════╝██╔════╝
██████╔╝█████╗  █████╗  ██╔██╗ ██║    ██║  ███╗██║   ██║█████╗  ███████╗
██╔═══╝ ██╔══╝  ██╔══╝  ██║╚██╗██║    ██║   ██║██║   ██║██╔══╝  ╚════██║
██║     ███████╗███████╗██║ ╚████║    ╚██████╔╝╚██████╔╝███████╗███████║
╚═╝     ╚══════╝╚══════╝╚═╝  ╚═══╝     ╚═════╝  ╚═════╝ ╚══════╝╚══════╝

██╗███╗   ██╗    ██╗   ██╗ █████╗  ██████╗ ███████╗███████╗███╗   ██╗
██║████╗  ██║    ██║   ██║██╔══██╗██╔════╝ ██╔════╝██╔════╝████╗  ██║
██║██╔██╗ ██║    ██║   ██║███████║██║  ███╗█████╗  █████╗  ██╔██╗ ██║
██║██║╚██╗██║    ╚██╗ ██╔╝██╔══██║██║   ██║██╔══╝  ██╔══╝  ██║╚██╗██║
██║██║ ╚████║     ╚████╔╝ ██║  ██║╚██████╔╝███████╗███████╗██║ ╚████║
╚═╝╚═╝  ╚═══╝      ╚═══╝  ╚═╝  ╚═╝ ╚═════╝ ╚══════╝╚══════╝╚═╝  ╚═══╝

Now go forth and create beautiful chaos! 🌈💀⚡

Name		Name	Last commit message	Last commit date
Latest commit History 9 Commits
.github/workflows		.github/workflows
assets		assets
static		static
templates		templates
.dockerignore		.dockerignore
.env.example		.env.example
.gitignore		.gitignore
Dockerfile		Dockerfile
LICENSE		LICENSE
README.md		README.md
chat.py		chat.py
cli_handler_simple.py		cli_handler_simple.py
file_readers.py		file_readers.py
main.py		main.py
ollama-chat-party		ollama-chat-party
rag.py		rag.py
requirements.txt		requirements.txt
shared_state.py		shared_state.py
web_server.py		web_server.py

License

psyb0t/ollama-chat-party

Folders and files

Latest commit

History

Repository files navigation