Activepieces

Activepieces is an open-source AI agent and workflow automation builder for business and technical teams, run in its cloud or self-hosted on your own servers.

Activepieces puts chat, AI agents, automated flows and data tables in one workspace connected to the apps a business already runs. The aim is that a whole team works with AI inside its existing tools instead of copying answers between windows. Behind it is an independent company in San Francisco, and its integration pieces are MIT-licensed open source. The same platform runs in the Activepieces cloud or on a company's own servers. Flows are drawn on a canvas, starting from a trigger such as Catch Webhook or a Human Input Web Form and running piece actions step by step. Agents get their own editor, where the builder picks a model from providers such as OpenAI, Anthropic, Gemini, Azure or Bedrock, plus the key that pays for it. Each agent then receives tools: piece actions, external MCP servers or one of the team's own flows. Knowledge files are split into chunks and indexed, so an agent searches them only when it needs something, and a Run Agent step can return typed fields instead of free text. Tables hold the data that flows produce, a built-in MCP server lets AI assistants call the same pieces, and interrupted runs resume from their last checkpoint. Operations, marketing, sales and support teams build their own automations here, while technical colleagues handle custom pieces and harder steps on the same platform. IT teams holding data on servers they control self-host it with Docker Compose or Helm. Software companies use Embed to place a white-label flow builder and agents inside their own product. A team can start on the free plan, move to a self-serve plan sized by a monthly credit allowance with extra credits charged as used, and reach enterprise or embedded contracts through sales. Support desks use it to triage incoming tickets, with an agent reading each one and posting urgent cases to a Slack channel. Sales teams enrich new leads before they land in HubSpot. Operations staff turn a web form submission into a clean row in Google Sheets, with an AI extraction step pulling out the fields in between. Because every agent sits inside a flow, a run can pause for a person to approve before anything is sent, and the run log shows each tool call afterwards. Activepieces fits teams that want AI agents acting on live business systems while keeping the option to run every flow on their own servers.