MCP runtime layer
Arcade is described as the MCP runtime between agents and the systems they need to reach, with auth, tools, and governance in one layer.
Arcade
Freemium
Arcade est un environnement d’exécution MCP pour les agents IA en production, qui associe autorisation des agents, outils fiables et gouvernance. Il aide les équipes à connecter les agents aux systèmes réels grâce à des déploiements flexibles et à des offres allant du gratuit à l’entreprise.
Qu'est-ce que Arcade ?
Arcade is an MCP runtime for production AI agents. It sits between agents and the systems they need to reach, combining secure agent authorization, reliable tools, and governance in one layer.
The product is aimed at teams that need agents to take real actions, not just generate text. Arcade’s site frames the core problem as moving past demo-stage agents by addressing authorization, tool reliability, and auditability together.
The pricing page shows a free Hobby plan, a paid Growth plan for teams moving agents to production, and an Enterprise option with custom pricing, dedicated infrastructure, and compliance-oriented controls such as audit logs, RBAC, and SSO/SAML.
The tools catalog shows production-ready connectors across common work systems such as Google Workspace, Slack, GitHub, HubSpot, Outlook, and other services, giving agents a path to execute actions in real environments.
Que peut faire Arcade ?
Agent authorization
Agents can act on behalf of real users with dynamic permissions, while credentials stay inside the runtime and existing identity providers plug in.
Agent-optimized tools
Arcade tools are presented as production-ready and agent-optimized, designed to map user intent to the right API call and avoid hallucinated parameters and retries.
Central governance and audit
The platform provides a central control plane for policy and audit, so every action can be traced by user and system without slowing delivery.
Stack flexibility
Arcade works with any LLM, framework, identity provider, or MCP client, which keeps the runtime flexible across existing stacks.
Broad tool catalog
The tools catalog shows a large set of available integrations across productivity, communication, developer tools, search, sales, and CRM workflows.
Cas d'utilisation
“User-authorized agent actions”
Use Arcade when an agent needs to act on behalf of a user in systems that require permissions, so the runtime can handle authorization without exposing shared credentials.
“Production tool execution”
Use it when agent tool calls need to be reliable enough for production workflows, especially where bad parameter mapping or retries would waste time or tokens.
“Governed agent operations”
Use it when teams need a central place to define policy and review what actions an agent took across systems, users, and tool calls.
“Business-system integrations”
Use it when you want to connect agents to common workplace systems such as Google Workspace, Slack, GitHub, HubSpot, Outlook, or similar tools from the catalog.
“Incremental stack adoption”
Use it when your team already has an LLM, framework, or identity provider and wants an MCP runtime that fits into the existing stack.
Questions fréquentes
What is Arcade?
Arcade is positioned as the runtime layer between AI agents and the systems they need to reach. The source describes it as combining agent authorization, reliable tools, and governance in one MCP runtime.
What does Arcade integrate with?
The source says Arcade works with any LLM, any framework, any identity provider, and any MCP client, so teams can fit it into existing agent stacks rather than rebuild around a new framework.
Does Arcade have free and enterprise options?
Arcade offers a free Hobby plan, a paid Growth plan, and an Enterprise plan with custom pricing and dedicated infrastructure. The pricing page also mentions a startup program for startups, nonprofits, educational institutions, and small companies under 100 employees.
What problems does Arcade focus on solving?
The product pages emphasize secure authorization, agent-optimized tools, and centralized governance. The pricing page also lists audit logs, RBAC, and SSO/SAML on Enterprise.
What kinds of systems can agents connect to through Arcade?
The tool catalog page shows production-ready tools across productivity, communication, developer tools, search, sales, and CRM categories, with examples such as Google Workspace apps, Slack, GitHub, HubSpot, and X.
Informations rapides
- Category
- MCP runtime for AI agents
- Website
- arcade.dev
- Pricing
- Free Hobby plan, paid Growth plan, Enterprise with custom pricing
- Primary users
- Teams building production AI agents
- Deployment
- Cloud, on-prem, air-gapped, or hybrid
- Tool coverage
- Production-ready tools across productivity, communication, developer tools, search, sales, and CRM
Analyse du trafic de Arcade
Les données de trafic sont fournies à titre indicatif uniquement.
- Évaluation du domaine
- 69
Alternatives à Arcade
Kastra
kastra.ai
Kastra est une infrastructure d’autorisation pour les systèmes d’IA qui vérifie les prompts, appels d’outils, commandes shell, requêtes API et actions du navigateur avant leur exécution. Elle aide les équipes à appliquer les politiques, conserver des journaux d’audit signés et gouverner les workflows d’IA locaux et d’entreprise.
Caveman
caveman.so
Caveman est une pile d’efficacité des tokens destinée aux agents IA et aux workflows de production. Elle aide les équipes à réduire le contexte éligible, à acheminer le trafic et à mesurer les économies vérifiées grâce à des outils locaux, un SDK TypeScript et une couche cloud gérée.
TinyFish
mino.ai
TinyFish est une plateforme d’entreprise pour agents web IA qui combine recherche en temps réel, récupération de pages, automatisation du navigateur et workflows en plusieurs étapes. Elle aide les équipes à exploiter à grande échelle des sites dynamiques ou authentifiés.
Lunen
lunen.ai
Lunen est un plan de contrôle gouverné pour créer et exploiter des agents d’IA sur les systèmes métier existants. Les équipes peuvent approuver leurs actions, gérer les autorisations et conserver un journal d’audit partagé en production.
Phinq
www.phinq.co
Phinq est une couche open source de gouvernance à l’exécution pour les agents IA. Elle intercepte leurs appels d’outils, les classe selon leur niveau de risque et suspend les actions les plus risquées pour obtenir une approbation humaine avant exécution.
MuleRun
mulerun.com
MuleRun est une plateforme d’agents IA pour automatiser les workflows de bout en bout : recherche, création de documents et présentations, production web et surveillance proactive. Plans gratuit, payants et entreprise disponibles.