Back to directory
Infrastructure Clouds

Arcade

MCP runtime for secure, production-ready AI agents with user-specific permissions

What makes Arcade different

Arcade is an MCP (Model Context Protocol) runtime specifically designed to make AI agents production-ready at enterprise scale. Unlike traditional cloud providers that focus on compute and storage, Arcade addresses the critical gap between AI models and real-world business systems by providing secure, user-specific authorization for AI agents.

The platform’s core innovation is enabling multi-user agents that act with individual user permissions rather than service accounts, integrating seamlessly with existing OAuth and identity provider flows. This approach solves the authentication and authorization challenges that prevent most AI agents from reaching production environments in enterprise settings.

Arcade offers the largest catalog of agent-optimized MCP tools—not just API wrappers—built specifically for AI agents to deliver better reliability and lower costs at scale. The platform includes complete visibility and governance controls, allowing organizations to deploy agents while maintaining compliance and security standards.

Pricing model

Arcade operates on a usage-based pricing model, though specific pricing tiers are not publicly disclosed. The platform appears to be in beta phase, with organizations able to join the beta program through their website. This suggests they’re likely working with early customers to establish pricing based on actual usage patterns and value delivery before announcing standard rates.

When it fits

• Enterprises needing AI agents to access multiple business systems with proper user authorization • Organizations requiring secure, compliant AI agent deployments with governance controls • Development teams building multi-user AI applications that need OAuth and IDP integration • Companies wanting to scale from AI pilots to enterprise-wide agent deployments • Teams needing flexible deployment options (cloud, VPC, on-premises, or air-gapped environments)

When it doesn’t

Arcade is not suitable for simple single-user AI applications or basic chatbots that don’t require complex business system integrations. It’s also not the right choice for organizations that need traditional cloud compute, storage, or infrastructure services rather than AI agent-specific capabilities.

Inclusion criteria

Arcade meets all 3 inclusion criteria: The website provides transparent information about their pricing approach (usage-based model), offers self-service signup through their “join the beta” program, and maintains public documentation and developer resources that indicate operational transparency.