crewAI

CrewAI is a platform for building and running AI agents inside a business. It covers the path from working out what to automate through to operating the result: a discovery step that reads tickets, chats, applications and workflows and returns automation opportunities ranked by effort, value and readiness; a builder that produces agent workflows from a prompt or a visual canvas; a runtime that executes them; and a control plane that governs every run. It targets two audiences at once, business teams who want to assemble automations without writing code and engineers who want a code first API and full control. Agents are organized by role, which keeps orchestration legible, and workflows can be made deterministic. The visual editor exports to Python, and a finished workflow can be exported as an MCP server or as a UI component. Governance includes real time tracing of every model call, tool call and memory read with cost accounting, role based access control, immutable audit trails, enterprise identity integration, human approval gates during execution, and runtime hooks that apply PII redaction and policy checks. Deployment runs on the vendor cloud, a dedicated VPC or customer infrastructure, and evaluation integrates with Arize, Galileo, Datadog and Patronus. Two plans are published. Basic is free and includes the visual editor and AI copilot, GitHub integration, two automations and 50 workflow executions per month on the vendor cloud. Enterprise is custom priced and adds single sign on with Microsoft Entra or Okta, role based access control, workload identity, PII redaction and policies, flexible overage on executions, a 45 day onboarding program and dedicated support channels. Forward deployed engineering and training are offered separately as paid additions.