This is a project that demonstrates how to use several AI technologies:
- OpenAI chat completion API
- Deepgram speech-to-text API
- Replicate API to run the miniGPT vision encoder model
- gpt-3-encoder package to tokenize strings into tokens (which are used to set limits on the length of prompts sent to OpenAI)
To see this project working, you can clone the project and then do the following.
npm install
Create a .env
file at the root of the project and then go into the .gitignore
file and add .env
to the list. This will make sure that the .env
file does not get pushed up to github if you choose to push the project up to github.
Add API keys to the .env
file as you see in the .env-example
file. Go to each of these websites to sign up for an API key:
npm run dev
node index.js
If you have any questions, you can reach out to me on twitter