LocalAI-Assistant

Private, Offline AI Chatbot powered by Ollama + Streamlit + Whisper + LLaVA + RAG

---

Overview

LocalAI-Assistant is a fully private, offline AI-powered assistant that runs directly on your local machine. It integrates multiple advanced AI models like Llama, Mistral, DeepSeek, Phi, TinyLlama, and LLaVA, providing a seamless and privacy-focused experience without any cloud dependency.

Key Features

• Chat with LLMs (Offline) — Converse with AI models locally
• PDF Summarization & Document Q&A — Upload documents and interact with them using AI
• Voice Input & Output — Convert speech to text (Whisper) and text to speech (pyttsx3) — fully offline
• Image Analysis with LLaVA — Understand and analyze image content through AI-powered vision models
• Chat with Documents via RAG — Retrieval-Augmented Generation for querying custom knowledge bases
• Multi-Chat Memory Management — Auto-save chats with options to rename, delete, and restore from a recycle bin
• 100% Local & Private — No internet required, no data leaves your machine

---

Tech Stack

Component	Description
Ollama	Run LLMs locally (Supports Llama3, Mistral, DeepSeek, Phi, TinyLlama, LLaVA)
Streamlit	Web-based user interface for easy interaction
LangChain	Enables document-based RAG (Retrieval-Augmented Generation)
Whisper (Offline)	Speech-to-Text model for voice input
pyttsx3 (Offline)	Text-to-Speech for voice responses
LLaVA	Vision-Language model for AI-powered image analysis
Local JSON Storage	Chat history, knowledge base, and recycle bin management

---

---

Installation

1️⃣ Install Ollama

• Download and install Ollama from 👉 https://ollama.com/download

• Run Ollama in the background:

ollama serve

2️⃣ Clone the Repository:

git clone https://github.com/your-username/LocalAI-Assistant.git
cd LocalAI-Assistant

3️⃣ Install Python Dependencies

Create a virtual environment (recommended):

python -m venv offenv
# Activate it:
offenv\Scripts\activate  # On Windows
source offenv/bin/activate  # On Mac/Linux

pip install -r requirements.txt

4️⃣ Pull Models via Ollama:

ollama pull llama3
ollama pull mistral
ollama pull deepseek-coder
ollama pull phi3
ollama pull tinyllama
ollama pull llava

▶️ How to Run

Run the app:

streamlit run app.py

Open in browser:

http://localhost:8501