Skip to content

Latest commit

 

History

History
70 lines (46 loc) · 2.23 KB

README.md

File metadata and controls

70 lines (46 loc) · 2.23 KB
Raw

GPT Functions with Encore

This is an example of using ChatGPT Functions with Encore.

ChatGPT functions allow developers to embed functions directly into the system prompt when interacting with the AI model. Function calling opens up several use cases, including:

  • Creating chatbots that leverage external/internal APIs
  • Converting natural language into API calls
  • Extracting structured data from text

In this example we are feeding in Encore APIs as the functions to the model. Because we are using Encore APIs we can see exactly which functions the model called in the generated trace data.

The gpt endpoint is this example could be interesting both for customers and for internal use. Customers could use it with prompts like "Recommend me a book that is similar to Kill a Mockingbird". For internal use, it could be used to quickly access data from the database with prompts like "Give me the book with the id of a1".

Example prompts:

  • Recommend me a book that is similar to Kill a Mockingbird
  • Give me the book with the id of a1
  • List all historical books
  • What is the book The Girl with the Dragon Tattoo about

Developing locally

When you have installed Encore, you can create a new Encore application and clone this example with this command.

encore app create gpt-functions-example --example=ts/gpt-functions

Running locally

You need to supply an OpenAI API Key to run this example:

encore secret set OpenAIAPIKey

Then run your application:

encore run

Local Development Dashboard

While encore run is running, open http://localhost:9400/ to access Encore's local developer dashboard.

You can make requests in the API explorer to the book.gpt endpoint to test the functionality.

Deployment

Deploy your application to a staging environment in Encore's free development cloud:

git add -A .
git commit -m 'Commit message'
git push encore

Then head over to the Cloud Dashboard to monitor your deployment and find your production URL.

From there you can also connect your own AWS or GCP account to use for deployment.

Now off you go into the clouds!