100% PRIVATE • ZERO TELEMETRY • LOCAL AI

Download CodAI Offline.

The open-source, local AI coding companion that runs 100% on your local hardware. Zero cloud dependencies, zero code harvesting.

v2.1.0 Latest Release
SHA-256 Signed
GPL-3.0 License
macOS icon
v2.1.0

macOS

Apple Silicon & Intel

System Requirements:
  • macOS 12.0 (Monterey) or later
  • 8GB RAM minimum (16GB+ recommended)
  • Metal GPU acceleration supported
Windows icon
v2.1.0

Windows

Windows 10/11 x64

System Requirements:
  • Windows 10/11 64-bit
  • 8GB RAM minimum (16GB+ recommended)
  • DirectX 12 or Vulkan GPU support
Linux icon
v2.1.0

Linux

AppImage & Universal

System Requirements:
  • Linux Kernel 5.10+
  • glibc 2.31 or later
  • Vulkan / ROCm / CUDA supported
Terminal Quick Install

Install via Command Line

View Installation Script
$ curl -fsSL https://codai.pro/install.sh | bash
PS> iwr -useb https://codai.pro/install.ps1 | iex

Supported Local AI Models

CodAI integrates with lightweight local LLMs optimized for hardware acceleration.

0.5B (4-bit Quant)

Qwen2.5-0.5B-Instruct

Ultra-Fast Local Completion

VRAM/RAM: ~500MB RAM/VRAM
1.0B Parameters

Llama-3.2-1B-Instruct

Meta Lightweight Instructor

VRAM/RAM: ~1GB RAM/VRAM
1.1B Parameters

tinyllama:1.1b

Ultra Compact Footprint

VRAM/RAM: ~1.1GB RAM
0.8B Parameters

Qwen 3.5-0.8B

Next-Gen Compact Reasoning

VRAM/RAM: ~800MB RAM/VRAM

Build Directly from Source

Inspect the transparent codebase, audit security guarantees, or compile CodAI directly using Node.js and Cargo on your local machine.