---
title: Quickstart | Lightcone
description: Go from zero to a working browser session in under 5 minutes.
---

## 1. Install the SDK

Terminal window

```
pip install tzafon
```

Terminal window

```
npm install @tzafon/lightcone
```

## 2. Get your API key

Sign up at [lightcone.ai](https://lightcone.ai) and copy your API key from the [dashboard](https://lightcone.ai/developer).

Set it as an environment variable:

Terminal window

```
export TZAFON_API_KEY=sk_your_api_key_here
```

Add this to your `.env` file so it persists across terminal sessions.

## 3. Create a browser and take a screenshot

quickstart.py

```
from tzafon import Lightcone


client = Lightcone()


with client.computer.create(kind="browser") as computer:
    computer.navigate("https://wikipedia.org")
    computer.wait(2)


    result = computer.screenshot()
    print(computer.get_screenshot_url(result))
    # The session terminates automatically when the block exits
```

quickstart.ts

```
import Lightcone from "@tzafon/lightcone";


const client = new Lightcone();
const computer = await client.computers.create({ kind: "browser" });


await client.computers.navigate(computer.id!, { url: "https://wikipedia.org" });


const result = await client.computers.screenshot(computer.id!);
console.log(result.result?.screenshot_url);


await client.computers.delete(computer.id!);
```

The TypeScript SDK returns optional types for fields like `id`. The non-null assertion (`!`) is safe here because `id` is always present after successful creation. You’ll see this pattern throughout the TypeScript examples.

The Python SDK has a high-level `client.computer.create()` wrapper with automatic cleanup via `with`. The TypeScript SDK also has a high-level `ComputerSession` class with the same convenience methods — see the [Computers](/guides/computers/index.md) page for details. The examples in these docs use the lower-level `client.computers` resource for explicitness.

Run it:

Terminal window

```
python your_script.py
```

Terminal window

```
npx tsx your_script.ts
```

You should see a screenshot URL printed to the console. Open it in your browser to see the captured page.

The output looks something like:

```
https://screenshots.tzafon.ai/abc123-def456.png
```

## What just happened?

1. **Created** a sandboxed Chromium browser session in the cloud
2. **Navigated** to Wikipedia
3. **Captured** a screenshot and got back a URL
4. **Terminated** the session (automatically in Python, explicitly in TypeScript)

The whole process takes about 3–5 seconds.

## Customize the session

You can configure the session when you create it:

```
with client.computer.create(
    kind="browser",
    timeout_seconds=3600,
    inactivity_timeout_seconds=120,
    display={"width": 1280, "height": 720, "scale": 1.0},
) as computer:
    # ...
```

```
const computer = await client.computers.create({
  kind: "browser",
  timeout_seconds: 3600,
  inactivity_timeout_seconds: 120,
  display: { width: 1280, height: 720, scale: 1.0 },
});
```

| Parameter                    | Description                                                        |
| ---------------------------- | ------------------------------------------------------------------ |
| `kind`                       | `"browser"` for web automation, `"desktop"` for full Linux desktop |
| `timeout_seconds`            | Maximum session lifetime (default: plan-dependent)                 |
| `inactivity_timeout_seconds` | Idle timeout before auto-termination (default: plan-dependent)     |
| `display`                    | Viewport dimensions and scale factor                               |
| `persistent`                 | Save cookies/storage for reuse across sessions (default: `false`)  |
| `use_advanced_proxy`         | Route through residential proxy for stealth (default: `false`)     |

## Next steps

- [**Computers**](/guides/computers/index.md) — understand sessions, lifecycle, and all available actions
- [**Automate a browser**](/guides/automate-a-browser/index.md) — end-to-end browser automation walkthrough
- [**Run an agent**](/guides/run-an-agent/index.md) — let AI complete tasks autonomously
- [**How Lightcone works**](/guides/how-lightcone-works/index.md) — architecture and the three API layers