https://github.com/imrany/insight
A voice artificial intelligence assistant
https://github.com/imrany/insight
artificial-intelligence gemini-api webspeech-api
Last synced: 2 months ago
JSON representation
A voice artificial intelligence assistant
- Host: GitHub
- URL: https://github.com/imrany/insight
- Owner: imrany
- License: mpl-2.0
- Created: 2024-11-02T09:04:40.000Z (7 months ago)
- Default Branch: main
- Last Pushed: 2024-11-17T17:08:42.000Z (6 months ago)
- Last Synced: 2025-01-19T23:20:25.253Z (4 months ago)
- Topics: artificial-intelligence, gemini-api, webspeech-api
- Language: TypeScript
- Homepage: https://insight-ai-ecru.vercel.app
- Size: 10.5 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
### Insight
Insight is an AI chatbot built in **Nextjs** that allows users to enter prompts using speech rather than text.
It uses **Gemini** under the hood to give back responses with the user can listen to or download a an audio file fo later.
### Why did we chose voice input only rather than both text and voice input
We've noticed a trend in AI chatbot like **ChatGPT** where users interacts with it mainly through text input, this can
limit people living with low or lack of vision and the elderly. Therefore we chose voice input for **insight**.
It can be used for assistance with general questions like *What's the recipe of chocolate cake...?* and other genral questions.
We've combine both speech and generative ai to create this program.
### Challenge statetement
We choose the open category of the hackathon.
#### Open category 🌟
Feeling adventurous? Build anything that combines two or more modalities—go wild with your ideas! Whether it's voice-controlled art generators or AI-powered virtual assistants, the only limit is your imagination.
### Features 🚀
- Speech recognition or input: Using `Web SpeechRecognition`
- Speech output:
Using `SpeechSynthesisUtterance` we convert text response to speech (Text-to-speech).
Using `MediaRecorder` we record it and store it for user to download.
Why this approach is not compute intensive and cost effective since some processes like text-to-speech is handled on the client side rathen than on the server-side (our side)
- Download audio files: Users would able to store and download audio files of their prompts and responses to later.
- Intelligent: We've integrate insight with **Gemini**, with this integration it can response fast and with nearly
accuracy to the subject matter.
### Demo video
You can view our demo on youtube, [https://youtu.be/zw7Cu6W9YCY](https://youtu.be/zw7Cu6W9YCY)
- Sign in page
- Home page
- How to download a chat as audio
- Short demo video
Your browser does not support the video tag.
### How to setup and run it locally
1. Ensure you have **Nodejs**, if not download here [https://nodejs.org/en/download/package-manager](https://nodejs.org/en/download/package-manager)
2. Clone this repository
```bash
git clone https://github.com/imrany/insight
```
3. Get into the project and open it with any IDE or code editor, is this we've used **vscode**
```bash
cd insight
code .
```
4. Install all project dependencies
```bash
npm install
```
5. Running the project
```bash
npm run dev
```
6. Open browser to [http://localhost:3000](http://localhost:3000)
7. You have to **Create an account** by clicking the `Get started` button
or you can sign in using my credentials
```
email: [email protected]
password: 12345678
```
8. Start by entering a prompt by pressing the `Microphone` button.
9. Once your voice prompt is recorded, you will receive a response which you can play or read by `double tapping` or `expanding` the card.
10. If you need any help, reach out here [[email protected]](mailto:[email protected])
### Live demo 🌏
Visit [https://insight-ai-ecru.vercel.app](https://insight-ai-ecru.vercel.app)