Open-source, secure environment with real-world tools for enterprise-grade agents.
Python agentaiai-agentai-agentscode-interpretercopilotdevelopmentdevtoolsgptgpt-4javascriptllmnextjsopenaipythonreactsoftwaretypescript
Update guide for custom sandboxes when using code interpreter SDK
Vasek Mlejnskya5dcc293e465
diff --git a/apps/docs/src/app/code-interpreter/template/page.mdx b/apps/docs/src/app/code-interpreter/template/page.mdxindex 41ed855ba..33ab5b1eb 100644--- a/apps/docs/src/app/code-interpreter/template/page.mdx+++ b/apps/docs/src/app/code-interpreter/template/page.mdx@@ -1,56 +1,59 @@import Link from 'next/link'import { FileCode } from 'lucide-react'-# Custom template using Code Interpreter--The Code Interpreter requires a template with a specific setup. If you want to build your own custom template and use Code Interpreter, you need to follow the instructions below.--Here's all the source code for the template:--<Link- className="w-fit flex items-center justify-start my-[-24px]"- href="https://github.com/e2b-dev/code-interpreter/blob/main/template"- rel="noopener noreferrer"- target="_blank"->- <FileCode- size={18}- className="text-brand-600"- />- <span>- Template source code- </span>-</Link>--## How to create custom template--If you want to build your own custom template and use Code Interpreter, you need to do:--1. Copy following files to your template folder: [`jupyter_server_config.py`](https://github.com/e2b-dev/code-interpreter/blob/main/template/jupyter_server_config.py), [`ipython_kernel_config.py`](https://github.com/e2b-dev/code-interpreter/blob/main/template/ipython_kernel_config.py) and [`start-up.sh`](https://github.com/e2b-dev/code-interpreter/blob/main/template/start-up.sh).-2. Setup jupyter server and kernel in your `Dockerfile` by adding these lines:-```docker-# Installs jupyter server and kernel-RUN pip install jupyter-server ipykernel ipython-RUN ipython kernel install --name "python3" --user--# Copies jupyter server, kernel config and start-up script-COPY ./jupyter_server_config.py /home/user/.jupyter/-COPY ./ipython_kernel_config.py /home/user/.ipython/-COPY ./start-up.sh /home/user/.jupyter/--# Sets permissions for the start-up script-RUN chmod +x /home/user/.jupyter/start-up.sh-```-3. Add the following option `-c "/home/user/.jupyter/start-up.sh"` to `e2b template build` command or add this line to your `e2b.toml`.-```yaml-start_cmd = "/home/user/.jupyter/start-up.sh"-```+# Using custom sandbox template & custom compute with Code Interpreter SDK++If you want to customize the Code Interprerter sandbox (e.g.: add a preinstalled package) you can do that by using a [custom sandbox template](https://e2b.dev/docs/sandbox/templates/overview).+++## Step-by-step guide+1. Create custom sandbox by following [this guide](https://e2b.dev/docs/guide/custom-sandbox)++2. Use prebuilt [E2B Code Interpreter image](https://hub.docker.com/r/e2bdev/code-interpreter) by replacing the `FROM` command in your `e2b.Dockerfile` with following++ ```Dockerfile+ FROM e2bdev/code-interpreter:latest+ ```++3. Copy [`start-up.sh`](./start-up.sh) to the same directory where's your `e2b.toml`-<Note>- Alternatively, you can use the [`e2bdev/code-interpreter`](https://hub.docker.com/r/e2bdev/code-interpreter) as base image.-</Note>+4. Run the following in the same directory where's your `e2b.toml`+ ```sh+ e2b template build -c "/home/user/.jupyter/start-up.sh"+ ```++5. Use your custom sandbox with Code Interpreter SDK++ **Python**+ ```python+ from e2b_code_interpreter import CodeInterpreter+ sandbox = CodeInterpreter(template="your-custom-sandbox-name")+ execution = sandbox.notebook.exec_cell("print('hello')")+ sandbox.close()++ # Or you can use `with` which handles closing the sandbox for you+ with CodeInterpreter(template="your-custom-sandbox-name") as sandbox:+ execution = sandbox.notebook.exec_cell("print('hello')")+ ```+++ **JavaScript/TypeScript**+ ```js+ import { CodeInterpreter } from '@e2b/code-interpreter'+ const sandbox = await CodeInterpreter.create({ template: 'your-custom-sandbox-name' })+ const execution = await sandbox.notebook.execCell('print("hello")')+ await sandbox.close()+ ```+++## Customize CPU & RAM of your sandbox+You can customize number of CPUs and MiB of RAM for your sandbox. To achieve that, specify the `--cpu-count` and `--memory-mb` options during the build step:+```sh+e2b template build -c "/home/user/.jupyter/start-up.sh" --cpu-count 4 --memory-mb 4096+```+The above will create a custom sandbox with 4 CPUs a 4 GiB of RAM.-## Install another kernels+## How to install another Python kernelsJupyter has ability to work with different than Python kernel. It even supports multiple kernels in one notebook. If you want to install another kernels.