Arcade

MCP actions runtime that authenticates, executes, and governs agent tool calls.

Visit site

Pricing

Free tier with monthly auth events and tool calls; Team from $25/mo plus usage; Enterprise custom.

Overview

Arcade is an actions runtime between ai agents and the systems they touch. It enforces identity and policy on every tool call, executes permission-aware actions, and leaves an audit trail of who did what. Teams use thousands of pre-built MCP tools plus custom tools, with cloud, VPC, on-prem, or air-gapped deployment. It works alongside common agent frameworks and IdPs rather than replacing them.

Key features

  • User-scoped auth via existing IdP and OAuth flows
  • 8,000+ permission-aware MCP tools plus custom tools
  • Pre- and post-call policy hooks
  • Single control plane for grant, revoke, and version tools
  • Action audit records streamable to SIEM
  • Framework-agnostic SDKs for Python and JavaScript
  • Cloud, VPC, on-prem, and air-gapped deploy options

Use cases

Show all 8Show fewer

Alternatives

  • Invicta AINo-code platform for building collaborative teams of tool-using ai agents.AI tool
  • MakeVisual AI automation platform for scenarios, agents, and 3,000+ app integrations.AI toolFreemium
  • Zapier AgentsAI teammates that act across 9,000+ Zapier apps with company knowledge and triggers.ai agentFreemium
  • Automation AnywhereEnterprise agentic process automation platform orchestrating ai agents, automations, and people.ai agent
  • CamundaOpen process orchestration platform coordinating ai agents, people, systems, policies, and approvals.ai agent
  • BotpressEnterprise AI agent platform for customer support tickets and multi-step workflows.AI toolFreemium

Last updated 2026-09-16