Tired of confusing official documentation searches? Our AI assistant, Billy, helps you find the answers you need instantly. Our solution curates your official documentation or any website into a searchable collection. You just need to spend some time for changing your website into a collection. Ask natural language questions and get the information you need directly - no more endless link clicking.
You can check the demo video here.
- Install GCloud CLI and check this link for setting up Vertext AI. (You need to have Google Cloud account and setup a project to use Vertex AI. If you don't, you can go to next step.)
- Get API Key for Google AI Studio. Google AI Studio is more straightforward and easy to use than Vertex AI above.
- Get the API Key token from Apify for scraping websites. Create an account and get the API key from here
- Create a .env file and add the credentials and API to it. Check the .env.example for example. Below is the example content inside the .env file.
# Gemini
GOOGLE_AI_STUDIO_API_KEY=xxxxxxxxxxxxxxxxxxxxxxxxxxxx
# vertex ai
PROJECT_ID=xxx-xxxx-xxx
LOCATION=us-central1
# Apify
APIFY_API_TOKEN=apify_api_xxxxxxxxxxxxxxxxxxxxx
conda create --name billy-bot python=3.10
pip install -r requirements.txt --no-cache-dir
You can download our sample collection, chroma_storage.tar.gz
cd billybot
wget https://github.com/cherish-noe/billybot/releases/download/v0.1/chroma_storage.tar.gz
Extract it.
tar xzvf chroma_storage.tar.gz
streamlit run app.py
Docker Installation only supports Google AI Studio based models at the moment.
./docker/build_docker.sh --build
docker run -it --rm -p 8511:8511 --name billy-bot billy-bot
- Vertex AI
- Google AI Studio
- Apify
- ChromaDB
- Streamlit