kagent

Local AI agent for all your tasks

kagent is a locally running AI agent designed for developers, researchers, and power users. It supports multiple local LLM backends, filesystem access, command execution, a knowledge base, and semi-autonomous operation with security controls.

---

Features

• Supports Ollama models

• Supports HuggingFace local models

• Automatic GPU detection (CUDA / ROCm / Metal)

• ReAct-style reasoning loop

• Full local filesystem interaction

• Shell command execution with output capture

• Web search & scraping

• Persistent local knowledge base (vector store)

• Risk detection + confirmation system

• Dynamic model switching

• Modular plugin architecture

• Cross-platform (Windows, Linux, macOS)

---

Requirements

• Python 3.11+
• ollama
• CUDA / ROCm / Metal drivers (if using GPU)

---

⚙️ Installation/ Setup

1. Fork the reop on your github.
2. Execute the command

git clone https://github.com/<your-username>/kagent.git

3. Move into the directory

cd kagent

4. Create virtual enviornment, and activate it.

python -m venv venv

# windows
.\venv\Scripts\activate

# linux / mac
source venv/bin/activate

5. Install dependencies

pip install -r requirements.txt

6. Download ollama and install it.
7. After ollama is successfully installed run ollama pull llama3, and verify it by ollama run llama3.

Alternatively, you can run the setup.py file.