activepieces: an open-source AI automation platform with a massive library of MCP servers for LLM integration

What it solves

Activepieces provides an open-source, self-hosted alternative to automation platforms like Zapier, allowing users to connect different software services and automate workflows without needing to write complex code.

How it works

It uses a "pieces" framework written in TypeScript, where each "piece" is an npm package representing an integration with a specific service (e.g., Google Sheets, OpenAI, Discord). Users can build automation flows using a no-code builder that supports logic like loops and branches. For AI-specific workflows, it includes native AI pieces and an AI SDK for creating agents. Additionally, all pieces are available as Model Context Protocol (MCP) servers, enabling them to be used directly with LLMs in tools like Claude Desktop, Cursor, or Windsurf.

Who it’s for

• Non-technical users who want to automate tasks using a visual no-code interface.
• Developers who want to build custom integrations using TypeScript or deploy a secure, self-hosted automation server.
• AI engineers looking to integrate LLMs with external tools via MCP servers.

Highlights

• MCP Toolkit: Over 280 pieces available as MCP servers for use with LLMs.
• AI-First: Native AI pieces and an AI SDK for building agents within flows.
• Extensible: Type-safe framework for creating new pieces as npm packages.
• Enterprise-Ready: Supports self-hosting, network-gapping for security, and human-in-the-loop approvals.