Experimentation repo for performing web based tasks using AutoGPT and Playwright.
Use the power of LLMs to run monotonus tasks in the browser and take screenshots using AutoGPT and Playwright.
Make sure you have bun installed and an OpenAI api key set
- Run
bun install - Give the agent a goal of what it should achieve together with a context
- Run with
bun run ./src/index.ts
The intended outcome is a group of screenshots from the flow the agent just ran using Playwright. However, it usually takes a few runs before it manages to run through the flow as intended.