https://github.com/trace2798/google_ai_hackathon
https://github.com/trace2798/google_ai_hackathon
Last synced: 3 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/trace2798/google_ai_hackathon
- Owner: trace2798
- License: mit
- Created: 2024-04-29T09:09:27.000Z (about 1 year ago)
- Default Branch: master
- Last Pushed: 2024-05-02T20:52:22.000Z (about 1 year ago)
- Last Synced: 2025-01-22T06:48:09.578Z (5 months ago)
- Language: TypeScript
- Homepage: https://google-ai-hackathon.vercel.app
- Size: 2.65 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Kenniscentrum: Where knowledge begins
This is a repository for my submission for Google AI Hackathon hosted on Devpost, which took place from March 18 until May 02, 2024.
Features:
- Implementation of Google Gemini API
- Langchain integration with gemini api
- Chat with your document
- Langchain wiki tool to search wikipedia and then generate response back using LLM
- Langchain agent integration with gemini AI to browser website and the internet.
- Implementation of server actions.
- Ability to soft/permanently delete thread and add custom prompt for thread.
- Restore soft deleted threads and files.
- Custom open source font.
### Prerequisites
**Node version 18.17.x** or higher. I am using v21.7.1 locally.
**API keys mentioned below and in the .env.example file**
### Cloning the repository
```shell
git clone https://github.com/trace2798/google_ai_hackathon.git
```
### Install packages
```shell
npm i
```
### Setup .env file
```js
# Using Clerk for authentication
NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY=
CLERK_SECRET_KEY=
# For this application I am using a MySql database from aiven.
DATABASE_URL=
NEXT_PUBLIC_APP_URL="http://localhost:3000"
# Both of them will have the same API key, langchain excepts it in GOOGLE_API_KEY wording.
GEMINI_API_KEY=
GOOGLE_API_KEY=
# using upstash vector as the vector database to store the embedded content of the document
UPSTASH_VECTOR_REST_URL=
UPSTASH_VECTOR_REST_TOKEN=
BRAVE_SEARCH_API_KEY=
```
### Start the run the development server:
```shell
npm run dev
```
Open http://localhost:3000 with your browser to see the result.