Skip to main content

​
Overview

Nango’s AI builder allows you to fully generate or assist in implementing Functions. You can use the IDE and models of your choice to generate and write Functions with minimal manual effort. AI is a game-changer for building many custom integrations fast, instead of having to use pre-built standard integrations provided by integration platforms. Nango’s fully code-based interface is designed to leverage advances in AI code generation. Unlike black-box integrations, Nango generates human-readable code that you can review, edit, and maintain—combining the speed of AI with the reliability and control of code you own.

​
How it works

Nango’s AI builder feeds the necessary context to the model of your choice. This includes:
  • All the knowledge about using Nango’s interface
  • Coding practices to write best-in-class integrations
  • Testing patterns for robust, production-ready code
The model handles the tedious process of researching and understanding the external API’s syntax and translates your integration use case into Nango Functions. This means you can easily use Nango’s solid infrastructure in combination with AI models for the business logic, mostly offloading the integration implementation and maintenance burden.

​
Example workflow

When using the AI builder, you’ll be prompted for the necessary information about your use case to generate the relevant Nango Functions:
  1. Describe your use case: Specify what data you need to sync or what action you want to perform
  2. Review generated code: The AI builder generates Functions with proper error handling, pagination, and rate limiting
  3. Test against live API: The AI builder tests the integration using connections you’ve set up on Nango
  4. Get production-ready code: Receive tested Functions ready to deploy

​
Capabilities

​
Automatic testing

The AI builder has instructions to test integrations against the external API using connections that you have set up on Nango. This gives you certainty that an integration fully works when the AI builder completes its task.

​
Mock-based test generation

By default, the AI builder generates tests for your Nango Functions, easily integratable in your CI/CD pipeline. These tests use mock API responses, so they run deterministically and don’t introduce noise in your CI/CD.

​
Battle-tested patterns

The AI Builder leverages years of experience building hundreds of integrations for Nango customers, ensuring generated code follows production best practices.

​
Supported models & interfaces

Nango’s AI builder works out of the box with:
  • Claude Code - Anthropic’s Claude via command line
  • Gemini CLI - Google’s Gemini models
  • Cursor - AI-powered IDE
  • Cursor CLI - Command line interface for Cursor
  • Cody CLI - Sourcegraph’s AI coding assistant
  • VS Code with Copilot - GitHub Copilot integration
  • Opencode - Open-source AI coding assistant
  • Droid by Factory - Factory’s AI development tool
  • And more…

​
Next step

Build with AI

Learn how to leverage AI agents to build integrations faster.