MiseAI is a powerful, private AI assistant built by Venti Labs. It runs 100% locally on your hardware — no cloud, no API keys, no data leaving your device.

Highlights

🧠 7B Parameters — Fine-tuned from Qwen 2.5 Coder 7B
🔒 Fully Private — Runs offline, no internet required after download
💻 Expert Coder — Production-ready code generation and refactoring
⚡ 8GB VRAM — Optimized to run on consumer GPUs
📦 GGUF Format — Ready for Ollama, llama.cpp, LM Studio

Quick Start (Ollama)

ollama run ventilabs/miseai

Or install the Venti CLI:

powershell

irm venti-labs.xyz/install | iex
venti launch mise

Model Details

Property	Value
Base Model	Qwen 2.5 Coder 7B
Fine-tuning	LoRA (QLoRA)
Quantization	Q8_0
File Size	~8.1 GB
Context Window	16,384 tokens
Max Output	8,192 tokens

Use Cases

Code Generation — Write production-ready code in any language
Code Refactoring — Optimize and restructure existing codebases
Problem Solving — Step-by-step reasoning through complex challenges
Technical Writing — Documentation, README files, and technical articles

ventilabs/MiseAI-1.1-GGUF

Venti MiseAI 1.1

Highlights

Quick Start (Ollama)

Model Details

Use Cases

Links

ventilabs/
MiseAI-1.1-GGUF