VibeFlow Chat 2.0 - Local AI Chat
A downloadable tool for Windows
Your AI. Your Rules. Your Hardware.
☁️ Forget cloud-based chats.
🧠 VibeFlow Chat 2.0 turns your PC into a powerful, private AI hub.
🤖 Run multiple LLMs at once, switch between them on the fly,
🌐 chat with your own documents, and access it all from any device on your network.
Remember when AI was about exploration and freedom?
🛠️ VibeFlow Chat 2.0 brings that spirit back. We rebuilt our chat from the ground up to create the ultimate personal AI hub that lives entirely on your machine.
🎯 VibeFlow Chat 2.0 isn't just a window to a single model; it's a command center for your entire local AI arsenal. juggling: Load up Llama, Mistral, and Gemma simultaneously.
⚔️ Pit them against each other in the same conversation.
📈 Feed them your own documents and get insights no cloud service could ever offer.
✨ With a sleek new interface, a powerful plugin system, and seamless access from any device on your Wi-Fi, this is the AI experience you've been waiting for.
🚫 No subscriptions, no data mining, no limits.
💪 Welcome to Beast Mode.
Features:
- Multi-Model Juggling: Load and run several different LLMs at the same time. Switch between them instantly within a single chat to compare responses and find the best tool for the job.
- Your Personal AI Server: Access your chat hub from your phone, tablet, or laptop — any device on your local network. Your conversations follow you, saved securely on your own machine.
- Sleek Pro UI: A completely redesigned interface built for power users. Enjoy full Markdown support, syntax highlighting for code, and easy one-click copy for developers.
- Future-Proof Plugin System: VibeFlow Chat is now a platform. The new plugin architecture allows for massive future expansion, from connecting to online services to adding custom local tools.
- 100% Private & Offline: Your models, your data, your conversations. Everything runs and is stored locally. No internet connection required, no Big Tech snooping on your data.
Download
Install instructions
VibeFlow Chat 2.0: User Guide
⚠️Heads up:
– The .exe is unsigned, so Windows might show a warning.
Click “More info” → “Run anyway” — you’re good.
– Firewall may ask for permission — allow it so the app can open your browser.
Step 1: Getting Started
1.1. Install the Engine: Ollama
VibeFlow Chat requires the Ollama service to function.
- Navigate to the official website: ollama.com.
- Download and install the version for your operating system.
- After installation, ensure the Ollama service is running (on Windows, you'll see the llama icon in your system tray).
1.2. Install & Run VibeFlowChat.exe
- Download the latest Windows release as a .zip archive from: vibe-coding-flow.com/downloads/vibeflow-chat-v-2-0/
- Extract the archive to a convenient folder.
- To Start: Run VibeFlowChat.exe. A console window with server logs will appear, and a browser tab will open automatically.
- To Stop: Close the black console window.
(Note: Linux and macOS versions are in development.)
1.3. The Console and LAN Access
- The console window shows all server activity and your LAN address.
- Example: http://192.168.1.105:8000
- Open this in any browser on the same Wi-Fi network to access VibeFlow Chat.
Step 2: Core Features
- Multi-Model Conversations: Switch between models in the same chat.
- Server-Side History: Stored in a local SQLite database.
- Nickname-Based Profiles: Retrieve chats by nickname.
- Full Chat Management: Create, load, or delete chats.
- Clean & Responsive UI: Optimized for desktop and mobile.
- Markdown & Code Highlighting: One-click copy.
Step 3: Recommended Models
For Standard PCs & Laptops (4-8GB VRAM)
- The All-Rounder: llama3:8b - Command: ollama pull llama3:8b - VRAM: ~5GB
- The Coder: codegemma:7b - Command: ollama pull codegemma:7b - VRAM: ~5GB
- The Super-Lightweight: gemma:2b - Command: ollama pull gemma:2b - VRAM: ~1.5GB
For High-End Workstations (16-24GB+ VRAM)
- The Pro Coder: deepseek-coder-v2 - Command: ollama pull deepseek-coder-v2 - VRAM: ~10GB
- The Heavyweight Conversationalist: deepseek-v2 - Command: ollama pull deepseek-v2 - VRAM: ~16GB
For "Monster" Rigs (48GB+ VRAM)
- The Ultimate Coder: deepseek-coder:33b - Command: ollama pull deepseek-coder:33b - VRAM: ~20-22GB
Step 4: Performance & Multi-User Support
By default, Ollama processes one request at a time.
- 8-16GB VRAM: Try OLLAMA_NUM_PARALLEL=2
- 24GB+ VRAM: Try 3 or 4
To configure (Windows):
- Create a system environment variable named OLLAMA_NUM_PARALLEL with your value.
- Restart the Ollama service.
Wishing you an excellent vibe while using VibeFlow Chat!
Leave a comment
Log in with itch.io to leave a comment.