e2b-dev/E2B — reverse-engineered prompt
Reverse engineered prompt
Build me an open source service that lets AI agents safely run code in temporary cloud sandboxes. A user should be able to sign up, get an API key, create a sandbox from Python or JavaScript, run shell commands inside it, read the output, and close it when they’re done.
Please include simple SDK examples for both languages, like creating a sandbox and running echo, plus a code interpreter style example where the user sends a short snippet of code and gets the result back. The main goal is that people can let an AI write and execute code without risking their own machine.
Make the sandbox isolated and secure by default, with clear places to plug in real cloud infrastructure later. Add basic docs, environment variable setup, and a self hosting guide that explains AWS and Google Cloud support at a high level. Keep the developer experience clean and easy to try locally. Look up current docs online if you need to.
Want more depth? Deep Reverse