Activepieces
Open source alternative to Zapier: a no-code workflow automation platform with 280+ TypeScript integrations, where every piece is also exposed as an MCP server for Claude Desktop, Cursor or Windsurf.
Overview
Activepieces is an open ecosystem alternative to Zapier: a visual builder where you connect 280+ services ("pieces") into automations, with loops, branches, retries, custom code steps and full versioning. Every piece is a TypeScript npm package, so contributors can build and publish new integrations, and 60% of the current library comes from the community.
How it works
Flows are built on a drag-and-drop canvas. Each step is a piece action or trigger; you can mix no-code steps with a code piece (JavaScript/TypeScript with npm) whenever the built-in pieces aren't enough. Because pieces are regular npm packages, local development supports hot reloading.
The distinctive feature for this directory: all 280+ pieces are also published as MCP tools, so the same integrations you use to build a flow can be handed directly to an LLM agent running in Claude Desktop, Cursor or Windsurf without rebuilding the integration layer twice.
- Self-host with Docker Compose (Postgres + Redis + app + worker containers).
- Build a flow visually, or expose selected pieces as an MCP server for an agent to call directly.
- Human-in-the-loop pieces let you pause a flow for approval or delayed execution before it continues.
Licensing note
The Community Edition is MIT. Content under packages/ee/ (enterprise
features) is licensed separately under Activepieces' own Commercial
License. That part isn't covered by the MIT core, so don't assume the
whole repository is permissive.
Examples
- Zapier-style SaaS-to-SaaS automations (Google Sheets, Slack, Discord, RSS, 200+ others) without writing integration code.
- Turning a piece (e.g. a CRM or database connector) into an MCP tool so a Claude Desktop or Cursor agent can call it directly.
- Long-running approval workflows using the "Chat Interface" or "Form Interface" human-input triggers.
Installation
You need Docker (with the Compose plugin) and Git. Nothing else runs locally: Postgres, Redis, the app and the worker all come up inside containers.
- Install Docker Desktop (or Docker Engine plus the Compose plugin) if you don't have it, and make sure it's running.
- Clone the repo and move into it:
git clone https://github.com/activepieces/activepieces cd activepieces - Start the stack:
Adddocker compose up-dto run it in the background. The first run pulls every image, so it can take a few minutes. - Once the containers are healthy, open the app in your browser and create the first admin account.
- To hand a piece to an agent instead of building a flow by hand, open that piece's MCP settings inside Activepieces and paste the generated MCP URL into Claude Desktop, Cursor or Windsurf's MCP configuration.
For reverse proxies, managed databases or Kubernetes, follow the full install guide linked in the setup command in the sidebar.
Related assets
Dify
Open-source LLM app development platform that combines a visual AI workflow canvas, RAG pipelines, agent tooling and model management in one workspace, for teams shipping production-ready AI apps.
Langflow
Visual Python platform for building and deploying AI agents and workflows, with full component source access, multi-agent orchestration and one-click export as an API or MCP server.
Graphify
Turns a codebase, plus its docs, PDFs, images, and video, into a queryable knowledge graph that a coding agent can traverse instead of grepping raw files, and never guesses silently: every edge is tagged EXTRACTED, INFERRED, or AMBIGUOUS so you know what was found versus inferred.