Create a QA chatbot for your data in a few minutes

A bot that knows your domain knowledge

We've just launched a new chatbot docker image that you can customize without writing code. All you need, is a document with information to search for answers in.
For example, you could load a PDF or a text document with your company's FAQs.
Or maybe an information guide about a product.

This is a quick guide to setting up your local environment.

First, create a directory for your project, and create an empty file in it named docker-compose.yml  

Inside the docker-compose.yml file, paste the following code -

    image: savantly/base-bot:latest
      - 9000:9000
      - ./data/docs:/app/data/docs
      - ./data/stores:/app/data/stores

In the project folder, create a directory called data

Inside the data directory, create a folder names docs and a folder named stores

In the data/docs folder, place a document containing the text you'd like the bot to reference when giving answers to questions.

Your project structure should look something like this now -


The application requires that you supply the OPENAI_API_KEY as an environment variable. So execute the following, and you insert your key -

export OPENAI_API_KEY=[use your key here]

Now run docker compose up and your chat bot will start and index the embeddings from your document.

Once the application starts, you can visit http://localhost9000

Check out the full code here -