Nango is an open-source platform for building product integrations that connects your application to 700+ APIs through a unified interface. It’s designed for engineering teams building SaaS products who need to integrate with external services like Google Calendar, Slack, or Salesforce without managing OAuth flows, token refresh, or scaling infrastructure manually. Nango solves the complexity of integration development by abstracting authentication and infrastructure while giving full control over integration logic via TypeScript.
Technically, Nango provides three primitives: Auth (managed OAuth/API keys), Proxy (authenticated API requests), and Functions (TypeScript-based integration logic). It supports deployment via CLI, CI/CD, and self-hosting on Kubernetes or Docker. The platform integrates with AI coding tools like Claude and Cursor, and works with LangChain, MCP, and any backend stack. Nango’s runtime handles retries, rate limits, tenant isolation, and observability with OpenTelemetry.
What You Get
- 700+ Pre-Built API Integrations - Out-of-the-box support for OAuth, API keys, and token refresh for APIs like Google Calendar, Slack, Salesforce, and Oracle with documented schemas and connection flows.
- AI-Generated Integration Code - Generate TypeScript syncs and actions from natural language prompts using AI tools like Claude, with full type safety and editable code output.
- Unified API Proxy - Make authenticated requests to any connected API through a single endpoint, with automatic credential injection, retry logic, and rate-limit handling.
- TypeScript Functions for Integrations - Write custom syncs, actions, and webhooks as TypeScript functions with built-in Nango SDK access to APIs, storage, and logging.
- Built-in Sync Engine - Schedule and run bidirectional data syncs (e.g., Google Calendar events) with pagination, batch saving, and delete tracking using nango.trackDeletesStart/End and nango.batchSave.
- Webhook Processing - Receive and route webhooks from any external API through a universal interface with built-in retry and delivery guarantees.
- Per-Customer Configuration - Customize integration behavior per tenant using dynamic configuration stored with each connection, enabling multi-tenant SaaS workflows.
- CLI & CI/CD Integration - Scaffold, test, and deploy integrations using the Nango CLI (nango create, dryrun, deploy) with full Git and CI/CD compatibility.
Common Use Cases
- AI Agent Tool Calling - An AI coding assistant uses Nango to call Google Calendar or Salesforce APIs on behalf of users, powered by MCP and LangChain integrations.
- SaaS Data Sync for RAG Pipelines - A startup syncs customer data from HubSpot and Zendesk into a vector database for AI-powered support automation, using Nango’s scheduled syncs and batch saving.
- Multi-Tenant API Unification - A B2B platform normalizes CRM data from 20+ providers into a single schema for its internal dashboard, using Nango’s unified API and per-customer config.
- Automated Customer Onboarding - A fintech company auto-creates bank connections via Plaid and syncs transaction data using Nango’s OAuth flows and webhooks to trigger compliance checks.
Under The Hood
Architecture
- Monorepo structure with over 20 isolated packages, each with clear dependency boundaries and TypeScript-based type safety
- Service-oriented design with distinct modules for HTTP routing, workflow orchestration, script execution, and event publishing
- Dependency injection via custom registries enabling dynamic swapping of database adapters and credential stores
- Event-driven communication powered by Redis pub/sub to decouple webhooks, data ingestion, and real-time sync triggers
- Configuration-driven provider system allowing dynamic integration loading through declarative YAML definitions
- Strict separation of concerns with shared type and utility packages enforcing contract-based inter-module interactions
Tech Stack
- Node.js backend with TypeScript, managed via monorepo workspaces and compiled with tsc
- PostgreSQL as primary datastore with Knex.js for type-safe migrations and queries, complemented by Redis for caching and pub/sub
- Multi-service architecture with dedicated packages for server, jobs, orchestrator, and runner, each with independent build and dev workflows
- Frontend applications built with React and Vite, tested with Vitest for unit and integration coverage
- Docker-based deployment with multi-stage builds and docker-compose orchestration
- Comprehensive tooling including ESLint, Prettier, Husky, and EditorConfig to enforce code consistency across the codebase
Code Quality
- Extensive test coverage spanning unit, integration, and end-to-end scenarios with robust state validation
- Modular design with well-defined interfaces and dependency injection enhancing testability and maintainability
- Structured error handling with consistent HTTP responses and comprehensive input/environment validation
- Strong TypeScript enforcement across APIs, database models, and webhooks ensuring type safety at every layer
- Consistent naming conventions and file patterns aligned with domain boundaries and test scope
- Strict linting and type-checking rules with no misuse of type assertions, ensuring API contract integrity
What Makes It Unique
- Unified connector layer abstracting 200+ API auth flows into a single standardized interface, eliminating custom integrations
- Native bidirectional sync engine with automatic schema evolution and conflict resolution, removing manual webhook management
- Embedded billing infrastructure that auto-provisions customers and payment methods directly from connection metadata
- Dynamic provider system enabling runtime addition of new integrations without code deployments
- End-to-end encrypted credential storage with per-environment key isolation and zero-knowledge access patterns
- Declarative sync configuration that auto-generates type-safe API clients, drastically reducing boilerplate