LLMSearch is a cross-platform web-searching application that fetches and processes multiple batches of search results from search engine APIs, summarizes the results using LLMs and presents the results to the user in the format of traditional search engines.
For project documentation, screenshots and reports, check out the project webpage wp2024.cs.hku.hk/fyp24051/.
Under frontend/, run npm i to install the necessary modules, then run npm run web to start the web development server.
Running the backend locally requires AWS SAM to be installed.
Under backend/main/, run npm i to install the necessary modules.
Then, run npm run dev to start the local development server.
Under frontend/, run npm run export, which runs npx expo export -p web.
The output directory is frontend/dist.
Set root directory to frontend/ and output directory to dist (frontend/dist).
Set build command to npm ci && npm run export.
Run sam deploy --guided to deploy the AWS SAM app to AWS.
- Creating backend stub
- Creating frontend prototype
- LLM-generated general summary and partial summary of search results
- Hosting backend on AWS
- Hosting frontend (on Cloudflare Pages)
- General summary chatbot function
- LLM-generated search queries and suggestions
- User authentication (signup and login)
- Evaluations of effectiveness done by LLM (LLM-as-a-judge)