Skip to content

lusikem/local-ai

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

32 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Local-AI. Offline Low-power AI for Everyone

Run light-weight AI models entirely on your own device, without cloud, subscriptions, or internet access.

Overview

Local-AI is an open-source project that empowers users to run light-weight language models (LLMs) on their local devices with limited computing power. Whether you're working with an older CPU-based laptop or deploying in remote or rural areas with unreliable connectivity, Local-AI gives you full control over your data and models.

This project uses the following stack:

Ollama — to run and customize LLMs on macOS, Linux, and Windows

Open Web UI — a browser-based chat interface, offline and open-source

Docker — for simple and reproducible deployments


Features

Run Offline — Works without internet access or cloud services

100% Private — Keeps all data on your device

Customizable — Fine-tune and personalize your own AI models

Free and Open Source — No subscriptions or vendor lock-in

Low Resource Compatible — Works on CPUs, no GPU required


🚀 Getting Started

1. Install Ollama

Download Ollama from: https://ollama.com/download
Or use terminal

Ollama supports: Mac (M1/M2/M3 optimized) Linux Windows (via WSL)

2. Run a Model

ollama pull mistral ollama run mistral

This starts a local REPL (interactive terminal). Here, you can type your prompts and interact with the model.

3. Launch Open Web UI (Optional GUI)

Use Docker to run Open Web UI locally:

Use Docker to run Open Web UI locally:

Once running, access the interface at: http://localhost:3000

4. Customize your model

You can fine-tune models like Mistral or LLaMA with your own datasets to adapt them to specific tasks, languages, or community knowledge.

Common Use Cases

  • Chatbots trained on regional or cultural knowledge (e.g., Swahili, Dholuo)
  • Domain-specific assistants (e.g., legal, medical, farming)
  • Q&A over PDFs, spreadsheets, or transcribed audio
  • Creative writing, translation, or grammar correction

Requirements

macOS/Linux/Windows

Docker (optional for GUI)

CPU with at least 8GB RAM is recommended

Terminal access

License

This project is released under the MIT License.

Contributing

Pull requests, forks, and improvements are welcome! Fork the repo Add your Swahili/other-language dataset Train and test locally Submit a PR from your fork to the original repo

Community & Purpose

This project is built for users/communities that have: Limited access to reliable internet Concerns over data privacy and AI surveillance A need for localized, culturally-relevant AI tools

By supporting open-source LLMs on local devices, we reclaim control of our digital futures.

About

Run AI models offline without relying on internet access or cloud infrastructure.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages