Roro AI is a real-time AI voice agent that helps users practice and improve their English communication skills by engaging in natural, spoken conversations with an intelligent assistant. Whether you're a student, fresher, or working professional, Roro AI provides a personalized and interactive way to build fluency, confidence, and clarity in spoken English.
🚀 Powered by Cutting-Edge Tech Roro AI combines a powerful stack of technologies to deliver seamless voice interactions:
🎧 MediaSoup – Real-time, low-latency audio communication
🧠 Gemini 2.0 Flash – Fast and context-aware large language model for generating smart, engaging responses
🗣️ Deepgram – High-performance speech-to-text (STT) and text-to-speech (TTS) engine for natural conversation flow
🎙️ Silero VAD – Lightweight and efficient voice activity detection for accurate speech segmentation and cleaner interactions
Roro AI listens intelligently, responds instantly, and adapts to your communication level — helping you improve through active speaking, not passive learning.
Here is a quick demo of the app. We hope you enjoy it.
Liked it? Please give a ⭐️ to Roro AI Repo.
Please access Roro AI using the URL:
Roro AI comes with a bundle of features already. You can do the followings with it,
- Engage in live, spoken conversations with an AI agent using low-latency audio powered by MediaSoup. Practice English just like you would with a human.
- With Silero Voice Activity Detection (VAD), Roro AI accurately detects when you're speaking — reducing noise, trimming silence, and keeping the conversation smooth and focused.
- Converse with an intelligent assistant backed by Gemini 2.0 Flash, a high-speed, context-aware language model that keeps the dialogue relevant and helpful.
- Your voice is transcribed in real time using Deepgram, ensuring fast and accurate understanding of what you say.
- Hear the AI speak back to you in a natural, human-like voice thanks to Deepgram's TTS, creating a realistic and immersive conversation experience.
- Clone the repository
git clone https://github.com/Daanish2003/roro-ai.git- Change the working directory
cd roro-ai- Install dependencies
pnpm install- Create
.envfile in root and add your variables
- Create .env file at
./apps/client/.env,./apps/server/.env,./apps/media/.envand./packages/database - add credentials from reference .env.example
- Docker setup
cd packages/database
docker compose up
- Run the app
pnpm run devThat's All!!! Now open localhost:3000 to see the app.
This project is licensed under the MIT License - see the LICENSE file for details.
Roro AI has all the potentials to grow further. Here are some of the upcoming features planned(not in any order),
- ✔️ AI Session Feedback
- ✔️ View Transcription
- ✔️ End of Utterence(Turn Detector)
If you find something is missing, Roro AI is listening. Please create a feature request from here.
- Incomplete Transcription sentence due to pause in between sentence which make forces the user to have short conversation sentence. This will generate response before even user intended to stop the speech
Solution: End of Utterence(Turn Detector)
Any kind of positive contribution is welcome! Please help us to grow by contributing to the project.
If you wish to contribute, you can work on any features listed here or create one on your own. After adding your code, please send us a Pull Request.
Please read
CONTRIBUTINGfor details on ourCODE OF CONDUCT, and the process for submitting pull requests to us.
We all need support and motivation. Roro AI is not an exception. If you found the app helpful, consider supporting us with a coffee.
