Activepieces

Activepieces is an open-source automation platform that lets teams build AI-powered workflows and agents without writing code. Founded by Ashraf Samhouri and launched in early 2023, it positions itself as an open-source alternative to Zapier and Make, with a community edition under the MIT license and enterprise features available under a commercial license. The GitHub repository has accumulated over 22,000 stars and the project reports around 270+ contributors.

What It Is

Activepieces is a no-code/low-code workflow automation tool built around a type-safe TypeScript "pieces" framework. Each integration — called a piece — is an npm package that can be contributed by the community and published to npmjs.com. The platform covers the full automation stack: trigger-based flows, multi-step logic with loops and branches, AI agent construction, human-in-the-loop approvals, and built-in data tables. According to the project's GitHub README, over 60% of pieces are contributed by the community, and all 280+ pieces are simultaneously available as MCP (Model Context Protocol) servers for use with LLMs in tools like Claude Desktop, Cursor, and Windsurf.

Open-Source Architecture and Deployment Model

The Community Edition is MIT-licensed and self-hostable via Docker, Helm, or any cloud provider. Enterprise features — including SSO, SCIM provisioning, advanced RBAC, audit logs, and piece access controls — live under a separate commercial license in the packages/ee/ directory. This split-license model means the core automation engine is freely available, while governance and compliance tooling requires a commercial arrangement. The platform supports both a fully managed cloud (with EU and US data regions, SOC 2 Type II and GDPR compliance) and self-hosted deployments for organizations with strict data residency requirements.

AI Agent and MCP Integration

Activepieces frames itself as an "AI Adoption Stack" — a platform for rolling out AI tooling across an entire organization, not just to technical teams. The builder includes native AI pieces for experimenting with multiple LLM providers, an AI SDK for building custom agents, and a chat interface trigger for conversational flows. The MCP toolkit, described in the README as the "largest open source MCP toolkit," exposes all pieces as MCP servers, enabling LLM-based agents to call real integrations directly. Flows support human approval steps, delay execution, and form-based input triggers.

Integration Breadth and Extensibility

The platform advertises 754+ integrations at the time of writing, spanning CRM (HubSpot, Salesforce), communication (Slack, Gmail, Microsoft Teams), databases (Snowflake, Google Sheets), project management (Jira, Notion), e-commerce (Shopify), HR (BambooHR, Okta), and AI providers (OpenAI). Developers can build new pieces using the TypeScript framework with hot reloading for local development. Pieces are versioned and published to npm automatically on contribution, making the ecosystem self-sustaining.

Enterprise Governance Features

The commercial tier adds IT oversight capabilities designed for large organizations: SAML 2.0 SSO with support for Okta, Microsoft Entra ID, Google Workspace, Auth0, and OneLogin; SCIM provisioning for automatic user and group sync; granular RBAC with role definitions across admin, builder, and viewer levels; global connections management; piece access controls; and full audit logs. The platform also includes an AI adoption analytics dashboard showing active flows, hours saved, adoption rates by team, and cost savings — framed as a tool for IT teams to demonstrate ROI on automation rollouts.

Update: Version 0.85.4

The latest release is version 0.85.4, published on June 17, 2026, with the repository showing active pushes as recently as June 20, 2026. The project maintains a high commit frequency (tracked weekly on the GitHub badge) and an open issues count of 539, consistent with an actively developed platform. The GitHub About description has been updated to emphasize AI agents, MCPs, and the ~400 MCP servers available for AI agent use, signaling a deliberate product pivot toward the agentic AI tooling market alongside its established workflow automation base.