Getting started

1. Get an API key

Sign in at ai2in.dev, open API Keys in the dashboard, and create a key. It looks like ai2in_live_…. Keep it secret.

2. Install

Python

pip install ai2in

JavaScript

npm install @ai2in/sdk

CLI

npm install -g @ai2in/cli

3. Run code

Python

from ai2in import Sandbox

# Creates a sandbox on construction; the context manager kills it on exit.
with Sandbox(api_key="ai2in_live_…") as sbx:
    execution = sbx.run_code("""
import pandas as pd
df = pd.DataFrame({"gst_lakh_cr": [1.87, 1.73, 1.95]})
df.describe()
""")
    print(execution.text)          # the main result's text
    print(execution.logs.stdout)   # captured stdout

JavaScript

import { Sandbox } from "@ai2in/sdk";

const sbx = await Sandbox.create({ apiKey: "ai2in_live_…" });
const execution = await sbx.runCode(`
import pandas as pd
df = pd.DataFrame({"gst_lakh_cr": [1.87, 1.73, 1.95]})
df.describe()
`);
console.log(execution.text);
await sbx.kill();

CLI

ai2in login --key ai2in_live_…
ai2in run -e "import pandas as pd; print(pd.DataFrame({'x':[1,2]}).describe())"

The kernel is stateful — variables persist across run_code calls in the same sandbox, just like a notebook.

4. Configuration

The SDKs default to https://api.ai2in.dev. Point them elsewhere (a self-hosted engine) with base_url / baseUrl, or the env vars the CLI reads:

export AI2IN_API_KEY=ai2in_live_…
export AI2IN_BASE_URL=https://api.ai2in.dev

Next steps

• Core concepts
• Running code — rich outputs, streaming
• Sandbox lifecycle — pause/resume, timeouts