A production-ready, multi-modal AI platform featuring Advanced Voice Mode, Deep Research (Grokipedia), and Neural Speed Diagnostics.
Live Demo โข Report Bug โข Request Feature
Framed captures of the live app (current UI).
AssistMe isn't just a chatbot; it's a comprehensive AI workspace designed for power users. Built on a modern React + FastAPI architecture, it integrates state-of-the-art LLMs (Gemini, Claude, GPT, Grok) with custom neural modules for specific tasks like deep research, image generation, and network diagnostics.
The latest update introduces a massively reorganized codebase, enterprise-grade architecture, and powerful new "applets" within the assistant.
A revolutionary "Search & Synthesis" engine that generates comprehensive, wiki-style articles on any topic.
- Deep Web Search: Aggregates data from multiple sources.
- AI Synthesis: Generates deep-dive articles with citations.
- Streaming Architecture: Real-time content generation with visible thought process.
Real-time, natural conversation with aggressively optimized latency.
- Live Integration: Zero-latency turn-taking.
- "Thinking" Filter: Scrubbing of internal chain-of-thought for pure speech output.
- Visual Feedback: Dynamic audio visualizers synchronized to speech.
A complete generative art workstation built right into the chat.
- Multi-Model Support: Flux, DALL-E, Stable Diffusion.
- Style Presets: One-click styles for Anime, Digital Art, Photorealism, and 3D Render.
- Smart Prompting: Automatic prompt enhancement for better results.
A professional-grade network diagnostic tool.
- Neural Diagnostics: Analyzes connection quality for specific use cases (Gaming, 4K Streaming).
- Granular Metrics: Measures Jitter, Bufferbloat, and Packet Loss alongside speed.
- Real-time Graphs: Visualization of TCP stream stability.
First-class support for Indian languages.
- 22+ Languages: Hindi, Marathi, Tamil, Telugu, Gujarati, and more.
- Script Handling: Correct rendering of complex Indic scripts.
- Cultural Nuance: Localized context understanding.
AssistMe follows a clean, modular architecture designed for scalability.
graph TD
User[User] --> Frontend[React + Vite Frontend]
Frontend --> |REST/WS| Backend[FastAPI Backend]
subgraph Frontend
Components[UI Components]
Services[Service Layer]
Context[State Management]
end
subgraph Backend
Router[API Router]
ServiceLogic[Business Logic]
Providers[LLM Integrations]
end
Backend --> OpenRouter[OpenRouter API]
Backend --> Gemini[Google Gemini API]
Backend --> Tavily[Tavily Search API]
| Directory | Purpose |
|---|---|
src/components/features |
Standalone feature modules (Voice, Speedtest, Grokipedia) |
src/services |
Centralized API wrappers and data fetching logic |
src/context |
Global state management (Theme, Auth) |
backend/app/routes |
API endpoint definitions |
backend/app/providers |
LLM provider abstraction layer |
- Node.js
- Python
- API Keys: OpenRouter (Required), Google Gemini & Tavily (Recommended)
-
Clone the repo
git clone https://github.com/mangeshraut712/AssistMe-VirtualAssistant.git cd AssistMe-VirtualAssistant -
Install Frontend
npm install
-
Install Backend
cd backend python -m venv venv source venv/bin/activate # Windows: venv\Scripts\activate pip install -r requirements.txt
Create a .env file in backend/:
OPENROUTER_API_KEY=sk-or-v1-...
GOOGLE_API_KEY=AIzaSy... # For Voice & Imagine
TAVILY_API_KEY=tvly-... # For GrokipediaDevelopment Mode (Concurrent):
# In Root
npm run dev
# In Backend/
python -m uvicorn app.main:app --reload --port 8001The Vite frontend is deployed to free GitHub Pages from GitHub Actions (Deploy GitHub Pages workflow). Production URL:
https://mangeshraut712.github.io/AssistMe-VirtualAssistant/
The old Vercel site (https://assist-me-virtual-assistant.vercel.app/) is paused/blocked and is no longer the production host. The repository does not overwrite third-party deployment statuses. If the Vercel GitHub App still posts a failing check, an administrator must disconnect it and update any required-status rule in GitHub repository settings.
GitHub Pages is static only. It does not run FastAPI, Node serverless functions, or environment secrets.
API features (chat, TTS, images, auth) require a separate process:
- Local / self-host: FastAPI in
backend/(uvicorn app.main:app --port 8001). Seebackend/README.md. - Former Vercel serverless: files under
api/were for Vercel and are not served by GitHub Pages. Seeapi/README.md.
Point the frontend at a backend URL you control (for example via env / settings) if you host the API elsewhere. There is no paid hosting in this setup.
- Framework: React, Vite
- Styling: Tailwind CSS
- Animation: Framer Motion
- Icons: Lucide React
- State: React Context + Hooks
- Framework: FastAPI
- Server: Uvicorn (ASGI)
- Validation: Pydantic
- Network: httpx, aiohttp
We welcome contributions! Please see our CONTRIBUTING.md for details on how to get started, our code of conduct, and submission guidelines.
- Fork the Project
- Create your Feature Branch (
git checkout -b feature/AmazingFeature) - Commit your Changes (
git commit -m 'Add some AmazingFeature') - Push to the Branch (
git push origin feature/AmazingFeature) - Open a Pull Request
Distributed under the MIT License. See LICENSE for more information.
Built with โค๏ธ by Mangesh Raut
GitHub Profile โข
Live Demo
flowchart LR
A["Browser"] --> B["Frontend App"]
B --> C["API / Serverless"]
C --> D["Deploy"]
High-level flow of the deployed web experience and supporting services.



