Activepieces is an open source alternative to Zapier and n8n that lets users create AI-powered automation workflows without code, while giving developers the ability to extend functionality via TypeScript-based pieces. It solves the problem of fragmented automation tools by unifying workflow automation, AI agent creation, and MCP server generation in a single platform. Built for both non-technical users and developers, it supports self-hosting, enterprise security, and community-driven integrations.
The platform is built with TypeScript and Node.js, uses a modular pieces architecture published to npm, and generates MCP servers automatically for LLM integration. It supports deployment via Docker, Helm, or cloud-hosted options with SOC 2 and GDPR compliance, and integrates with over 200 services including OpenAI, Slack, Notion, and Gmail.
What You Get
- 200+ Integrations - Pre-built pieces for Gmail, OpenAI, Slack, Notion, HubSpot, and more, all open source and published to npmjs.com.
- MCP Server Generation - Every piece automatically becomes an MCP server usable with Claude Desktop, Cursor, and Windsurf for LLM agent integration.
- TypeScript Pieces Framework - Developers can create custom integrations using TypeScript with hot reloading for local development and npm publishing.
- No-Code Workflow Builder - Visual drag-and-drop interface with loops, branches, auto-retries, HTTP requests, and code blocks with AI-assisted data cleaning.
- Human-in-the-Loop Triggers - Built-in Chat Interface and Form Interface to pause workflows and require human approval or input before proceeding.
- Enterprise Security & Self-Hosting - Deploy on your own infrastructure with full data control, SAML 2.0 SSO, SCIM provisioning, and granular RBAC for user permissions.
Common Use Cases
- Building AI Agents for Sales Teams - A sales director uses Activepieces to create a Lead Scorer agent that pulls data from CRM, enriches with OpenAI, and auto-assigns leads based on scoring rules.
- Automating Customer Support Responses - A support team deploys an AI Support Agent that reads incoming emails, summarizes them with LLMs, and suggests responses using Notion and Gmail integrations.
- Internal Operations Automation - An operations manager builds a Daily Report Agent that pulls data from Google Sheets, Slack, and Jira to generate and send daily team summaries.
- Enterprise Workflow Governance - An IT team deploys Activepieces self-hosted with SAML SSO and SCIM to give marketing and sales teams automation power while maintaining compliance and access control.
Under The Hood
Architecture
- Monorepo structure with clearly delineated packages for engine, API, worker, and UI, enforced via TypeScript paths to prevent circular dependencies
- Event-driven communication between frontend and backend through named client events, enabling loose coupling and scalable state management
- Plugin-based piece system allows dynamic loading of community and core integrations without recompilation, fostering extensibility
- Dependency injection via TypeORM and service registries ensures clean separation of execution context and business logic
- Dockerized multi-stage builds optimize production deployments by stripping unused components and isolating runtime dependencies
- Vue-based UI components follow component-based patterns with centralized state management, decoupling presentation from business rules
Tech Stack
- Node.js backend powered by Bun for fast builds and isolated-vm for secure, sandboxed piece execution
- PostgreSQL and Redis form the persistent and caching layers, orchestrated through Docker Compose with volume persistence
- TypeORM manages database modeling and migrations with integrated CI/CD validation for schema integrity
- i18next and Crowdin enable comprehensive localization, while esbuild ensures efficient frontend bundling
- Comprehensive testing suite with Vitest and Playwright, backed by strict TypeScript typing and ESLint enforcement
Code Quality
- Extensive test coverage across unit, integration, and end-to-end scenarios, with consistent mock contexts for reliable isolation
- Strong type safety enforced throughout the codebase using TypeScript interfaces and literal types to eliminate runtime errors
- Clear error handling with explicit exceptions for edge cases in file and PDF operations, ensuring predictable user feedback
- Consistent naming, test structure, and configuration patterns across all modules, enhancing maintainability in a distributed codebase
- Standardized linting and build configurations ensure uniform quality standards across teams and packages
What Makes It Unique
- Native Scrapeless AI integration eliminates infrastructure overhead for web scraping, offering a seamless data extraction experience
- Real-time trigger health dashboards provide visual, time-based operational insights directly within the UI
- Context-aware tag inputs with dynamic validation enable intelligent user input handling without external dependencies
- Granular authentication flows with feature-flagged behavior adapt to edition-specific requirements seamlessly
- Community pieces are first-class citizens in the flow editor, creating a self-sustaining ecosystem of integrations
- Embedded flow telemetry transforms operational metrics into interactive visualizations, turning logs into actionable intelligence