Generate

Defang includes an AI-driven agent that translates natural language prompts to an outline for your project that you can then refine. The AI agent is available through the Defang CLI.

We are working on expanding the range of supported prompts and improving the accuracy of the AI agent. If you have any feedback or suggestions, please let us know by opening an issue.

info

This page is about Defang's AI tool for project outline generation. We also have an AI tool for debugging.

Example Prompts

Here are some example prompts that the AI agent can help you with:

A basic service with 2 REST endpoints. The default endpoint will be for health check and should return a JSON object like this: { "status": "OK" }. The /echo endpoint will echo back all request parameters in the response.

A simple service that runs in the cloud and serves a web page

A simple flask app with one route that responds with a json object

A simple next.js app that serves a static page

A simple api with a single route built on express

An HTTP GET endpoint that returns an HTML with a random meme from the imgflip API

An example of a more complex prompt

A service that uses bullmq and redis. The compose file should also include a redis service. The code should set up a queue. It should also expose three endpoints with express: one to add a job to the queue, another to check how many tasks are in the queue, and a last one that runs on / which is just a healthcheck. The code should also include a worker that processes the queue by logging what's in each task. Use environment variables to configure the redis connection. No password.