Build on Nativ.

Start with a complete loop.

Let a person select content, choose a language, review a localised result and apply it. Your host application handles selection and placement. Nativ supplies translation and media capabilities with the context in your workspace.

1. Set up the workspace. Create a Nativ account, configure target languages and add your brand voice, glossary and style guides.
2. Create an API key. Open Settings → API Keys. Keep shared application keys on your server.
3. Localise one small item. Use an SDK example below, then build a preview around the returned result. Usage requires available credits.
4. Apply after review. Keep the source identity with each request and check it has not changed before inserting a result.

TypeScript & Node.js

Install nativ-sdk and set NATIV_API_KEY in your server environment. This example prints a translation and review information.

npm install nativ-sdk

import { Nativ } from "nativ-sdk";

// Run on your server. Set NATIV_API_KEY in its environment.
const client = new Nativ({ apiKey: process.env.NATIV_API_KEY });

const result = await client.translate(
  "Make room for something new.",
  "French",
  {
    context: "Short campaign headline for a homeware brand",
    backtranslate: true,
  },
);

console.log(result.translatedText);
console.log(result.backtranslation);
console.log(result.rationale);

Read the TypeScript SDK documentation ↗

Python

The Python client supports synchronous and asynchronous use. The SDK also includes a command-line tool for shell workflows.

pip install nativ

from nativ import Nativ

# Set NATIV_API_KEY in your server environment.
client = Nativ()

result = client.translate(
    "Make room for something new.",
    target_language="French",
    context="Short campaign headline for a homeware brand",
    backtranslate=True,
)

print(result.translated_text)
print(result.backtranslation)
print(result.rationale)

Read the Python SDK documentation ↗

Choose your connection.

• SDKs: convenient clients for your application or backend.
• REST API: direct HTTP from other languages and environments. Check the API reference for request schemas.
• MCP: discoverable tools for AI assistants. Use the MCP connection guide.
• JSON tools: JSON requests with HTTPS media URLs for connector builders. Import the tools schema.

Make the integration worth using.

Show the original alongside the proposed result. Preserve successful items in a batch when another item fails. Make the retry target clear, and keep approval visible. These small product decisions turn an API call into a workflow people can trust.

Read the guide to building a complete plugin workflow →