The [Modal Vibe repo](https://github.com/modal-labs/modal-vibe) demonstrates how you can build
a scalable AI coding platform on Modal.
Users of the application can prompt an LLM to create sandboxed applications that service React through a UI.
Each application lives on a [Modal Sandbox](https://modal.com/docs/guide/sandboxes)
and contains a webserver accessible through
[Modal Tunnels](https://modal.com/docs/guide/tunnels).
For a high-level overview of Modal Vibe, including performance numbers and why they matter, see
[the accompanying blog post](https://modal.com/blog/modal-vibe).
For details on the implementation, read on.
## How it's structured
* `main.py` is the entrypoint that runs the FastAPI controller that serves the web app and manages the sandbox apps.
* `core` contains the logic for `SandboxApp` model and LLM logic.
* `sandbox` contains a small HTTP server that gets put inside every Sandbox that's created, as well as some sandbox lifecycle management code.
* `web` contains the Modal Vibe website that users see and interact with, as well as the api server that manages Sandboxes.
## How to run
First, set up the local environment:
```bash
python3 -m venv venv && source venv/bin/activate && pip install -r requirements.dev.txt
```
### Deploy
To deploy to Modal, copy `.env.example` to a file called `.env` and add your `ANTHROPIC_API_KEY`.
Also, create a [Modal Secret](https://modal.com/docs/guide/secrets) called `anthropic-secret` so our applications can access it.
Then, deploy the application with Modal:
```bash
modal deploy -m main
```
### Local Development
Run a load test:
```bash
modal run main.py::create_app_loadtest_function --num-apps 10
```
Delete a sandbox:
```bash
modal run main.py::delete_sandbox_admin_function --app-id
```
Run an example sandbox HTTP server:
```bash
python -m sandbox.server
```