https://github.com/20mup/ai-receptionist
Voice-enabled AI receptionist powered by Whisper, LangChain, and ElevenLabs — trained on company-specific data scraped from Systems Limited.
https://github.com/20mup/ai-receptionist
ai elevenlabs internship langchain streamlit whisper
Last synced: 2 months ago
JSON representation
Voice-enabled AI receptionist powered by Whisper, LangChain, and ElevenLabs — trained on company-specific data scraped from Systems Limited.
- Host: GitHub
- URL: https://github.com/20mup/ai-receptionist
- Owner: 20mup
- Created: 2025-06-23T18:38:04.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-06-23T19:57:59.000Z (about 1 year ago)
- Last Synced: 2025-06-23T20:34:30.522Z (about 1 year ago)
- Topics: ai, elevenlabs, internship, langchain, streamlit, whisper
- Language: Python
- Homepage:
- Size: 1.27 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# 🤖 AIVA — AI Voice Assistant for Enterprises
> Voice-powered receptionist built with Whisper, LangChain, FAISS, and ElevenLabs — built during my internship at Systems Limited.
AIVA (AI Voice Assistant) is a voice-interactive AI receptionist that can answer company-specific questions by listening to your voice, retrieving accurate info from real webpages, and responding with natural speech.
---
## 📺 Demo
[](https://youtu.be/_fyLJ0vlOlo)
---
## 🚀 Why I Built This
During my internship at Systems Limited, I saw an opportunity to bring generative AI to a real business use case: reception and onboarding. I pitched AIVA as a solo project to my manager and VP — and after 4 weeks of development, I presented it to 60+ employees and the President of IT.
---
## ✨ Features
- 🎙️ Voice-to-voice interaction (ask a question, hear a reply)
- 📚 Trained on 100+ webpages scraped from systemsltd.com
- 🧠 LangChain + GPT-3.5 for accurate retrieval and context
- 🗣️ ElevenLabs voice output with natural tone
- 💻 Built in Streamlit — no setup, just run and speak
---
## 🧰 Tech Stack
| Purpose | Tools Used |
|-------------------|------------------------------------|
| Speech-to-Text | [Whisper](https://github.com/openai/whisper) (OpenAI) |
| Knowledge Retrieval | FAISS + LangChain |
| Text-to-Speech | [ElevenLabs](https://www.elevenlabs.io) |
| Interface | Streamlit |
| Language Model | GPT-3.5 (via LangChain) |
| Web Scraping | BeautifulSoup |
---
## 🧠 How It Works
1. **Scrape company content** → chunk into documents
2. **Embed content using OpenAI embeddings + FAISS index**
3. **Capture voice via Whisper**, convert to text
4. **Query LangChain’s retrieval chain** → generate response
5. **Use ElevenLabs** to speak the result out loud
6. **Streamlit UI** glues the entire experience together
---
## 📁 Project Structure
```bash
ai-receptionist/
├── app/
│ ├── main.py # Streamlit UI
│ ├── chatbot.py # LangChain logic
│ ├── audio.py # Whisper + ElevenLabs integration
│ ├── st_custom_components/ # Custom JS audio recorder
│ └── utils/
│ ├── scraper_example.py
│ └── concatenate.py
├── new_faiss_index/ # Saved FAISS vector index
├── requirements.txt
├── .env.example
└── README.md
```
---
## 📝 Setup Instructions
1. **Clone the repo**
```bash
git clone https://github.com/20mup/ai-receptionist
cd ai-receptionist
```
2. **Create a virtual environment**
```bash
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
```
3. **Install dependencies**
```bash
pip install -r requirements.txt
```
4. **Add your API keys to `.env`**
```
OPENAI_API_KEY=your-openai-key
ELEVENLABS_API_KEY=your-elevenlabs-key
```
5. **Run the app**
```bash
streamlit run app/main.py
```
---
## 🤝 Credits
Developed by [Mousa Pirzada](https://www.linkedin.com/in/mousa-pirzada/)
Internship: Systems Limited — Cloud & Infrastructure Division
Special thanks to my mentors and VP for trusting me with this initiative.
---
> _From static FAQs to fluent conversations — AIVA makes companies sound smarter._