Code Interpreter on steroids for ChatGPT (by e2b)

Powered by E2B

AI Playgrounds

Click on the image to watch the demo

The E2B plugin for ChatGPT offers a code interpreter on steroids. It gives your ChatGPT instance access to a full sandboxed cloud environment. This environment includes:

• Access to Linux OS
• Ability to install programs
• File system operations (create, list, and delete files and directories)
• Run processes
• Internet access

These cloud instances serve as playgrounds for agents, enabling them to operate freely. The plugin is powered by the E2B API. To gain early access, join our Discord.

What can I do with this plugin?

This plugin provides three simple commands, each with exciting potential:

• RunCommand: Executes any shell command.
• ReadFile: Reads a file from a specified path.
• WriteFile: Writes content to a file at a specified path.

With these commands, you can:

• Run any programming language, including NodeJS, Go, Bash, Rust, Python3, PHP, Java, Perl, or .NET.
• Install headless Chrome and perform various web scraping tasks.
• Launch databases and servers.
• Execute terminal commands and create long-running processes.
• Deploy websites and install programs via terminal.

Installation

There are two routes to install the plugin:

1. Wait for OpenAI to approve the plugin for their store.
2. Have developer access to ChatGPT plugins and install the plugin by following the instructions below.

How to run the plugin locally

To install the required packages for this plugin, run the following command:

npm install

To run the plugin, you will need an API Key. Click here to get your API key. Then enter the next command:

E2B_API_KEY=*** npm run dev

Once the local server is running, follow these steps:

1. Navigate to chat.openai.com.
2. In the Model drop down, select Plugins (you won’t see this option if you don’t have access yet).
3. Select the Plugin store.
4. Select “Develop your own plugin.”
5. Enter localhost:3000, as this is where the server runs locally, and select “Find manifest file.”

Usage Examples

Here are a couple of practical applications:

• Install youtube-dl and use it to download a specific video from YouTube here.
• Start an HTTP server on port 3000.
• Clone this repository here, fix any typos in the readme, and push the changes.

How to upload/download files

The official ChatGPT Code Interpreter supports file uploads and downloads, but the E2B code interpreter doesn’t support this functionality natively (yet). However, you can work around this by using the curl or wget commands with services like S3 buckets:

Uploading your files to the plugin

1. Get an S3 bucket (or an alternative).
2. Upload your files there and make them public.
3. Tell ChatGPT to download those files using curl.

Downloading files from the plugin

1. Instruct ChatGPT to upload its files to the S3 bucket using curl.

What is E2B?

The E2B company is behind this plugin. They are creating an operating system for AI agents, featuring a set of low-level APIs. This infrastructure includes debugging, authorization, monitoring, and more, along with sandboxed cloud environments for agents to explore without limitations.

Development

To install dependencies, run:

npm install

Then start the reloading server by running:

npm run dev

API Routes

Tsoa is employed to generate the OpenAPI spec and create the server route boilerplate. It uses TypeScript decorators to define the API. You can edit the Controllers in src/plugin to modify the exposed API.

Documentation

The OpenAPI spec documentation is generated from JSDoc comments in the Controllers. Visit the tsoa docs for more information. To change the info section in the OpenAPI spec, you’ll need to modify the reformatting script rather than editing the openapi.yaml file directly.

Manifest

You may also want to edit the .well-known/ai-plugin.json file to change the metadata about the plugin.

Troubleshooting

If you encounter any issues, here are some common troubleshooting steps:

• Ensure you have the correct API Key and that it is set in the environment variable.
• Check if your local server is running at the correct port (3000).
• Make sure you have all required dependencies installed correctly.
• If you have additional questions or need support, check out fxis.ai.

For more insights, updates, or to collaborate on AI development projects, stay connected with fxis.ai.

At fxis.ai, we believe that such advancements are crucial for the future of AI, as they enable more comprehensive and effective solutions. Our team is continually exploring new methodologies to push the envelope in artificial intelligence, ensuring that our clients benefit from the latest technological innovations.