Activepieces
Self-hosted AI workflow automation with a visual builder and MCP tools. Runs on Docker or Kubernetes; cloud AI connections use your provider credentials.
Activepieces is a self-hosted automation platform for teams that want AI agents and app workflows on their own servers. It runs with Docker or Kubernetes, alongside a managed cloud offering. Self-hosting controls where the app runs; connections to Anthropic, OpenAI, AWS Bedrock or Google Gemini still send model requests to those providers using your credentials.
Teams can describe recurring work in plain language and edit the resulting flow in a visual builder. Agents choose tools from connected apps and MCP servers, with human approval before sensitive actions. Workflows support branching, retries, subflows and error handling, and developers can add TypeScript for custom logic. Interrupted runs resume at saved checkpoints.
The same app connections serve workflows and agents. Integrations include Slack, HubSpot and Google Sheets, while REST and GraphQL support covers other APIs. Claude and Cursor can call workflows through its MCP server; integration pieces also work as MCP tools in Claude Desktop and Windsurf. Shared tables let flows and agents read and update workspace data, and a knowledge base makes documents searchable by agents.
The Community Edition is open source under the MIT license, as are the integration pieces. Enterprise features use a commercial license. Developers can extend the app connections through its TypeScript framework. Activepieces also supports deployments with no outbound network access, but connections to cloud apps and remote AI providers need access to those external services.
