Enterprise teams evaluating AI agent governance platforms face a fundamental challenge: understanding what they will actually spend before committing to a vendor. Runlayer, one of the prominent players in the MCP gateway space, operates with zero public pricing, requiring every prospective customer to book a demo before learning what the platform costs. This opacity creates procurement friction, extended sales cycles, and budget uncertainty that platform engineering and IT security teams must navigate carefully.

This article breaks down what is known about Runlayer's pricing structure, identifies which services require custom quotes, analyzes cost estimation approaches teams can use, and compares this model against alternatives for enterprise AI agent governance.

Key Takeaways

• Runlayer publishes no pricing tiers, no "starts at" numbers, and no self-serve signup, forcing all teams through custom quote processes regardless of deployment size
• CEO Andy Berman has stated Runlayer uses a platform-fee model rather than per-user pricing, scoped to deployment size and required capabilities, though no dollar amounts are publicly disclosed
• Runlayer does not publicly disclose typical customer contract values; AWS Marketplace pricing dimensions should not be treated as evidence of what customers actually pay
• Total cost depends on deployment model: self-hosted environments add customer-operated infrastructure and maintenance, while Runlayer also supports managed infrastructure for MCP deployments
• Mintive publishes team-size bands and uses per-user licensing for active AI-agent users, but buyers still need a custom quote for dollar pricing
• Migration timing depends on connector inventory, identity configuration, policy translation, deployment model, and client setup; MCP compatibility can reduce protocol-level changes but does not eliminate platform-specific migration work

Understanding Runlayer's 2026 Pricing Models

What Runlayer Discloses (and Does Not Disclose)

Runlayer's pricing structure remains one of the least transparent in the AI agent governance market. The company does not publish:

• Standard tier pricing or feature breakdowns
• Per-user, per-seat, or per-agent rate cards
• Self-serve trial options or freemium tiers
• "Starting at" price points for budget estimation

Every team seeking to evaluate Runlayer must book a demo with their sales team to discuss packaging and deployment options. This approach contrasts with SaaS pricing norms where companies publish at least baseline pricing to help buyers qualify themselves.

What the Market Data Reveals

Several data points provide partial visibility into Runlayer's pricing reality:

Public Contract Data:

• Runlayer does not publicly disclose typical contract values or average contract value
• Marketplace pricing dimensions and contract ceilings do not establish what individual customers actually pay
• Buyers should request an actual quote rather than use Marketplace maximums as pricing benchmarks

Pricing Model Structure:

• CEO Andy Berman has said Runlayer uses a platform-fee model rather than per-user pricing
• He described pricing as scoped to deployment size and the capabilities required
• Runlayer does not publish dollar rates or standard quote ranges

The Procurement Friction Problem

For IT leaders and platform engineering teams, pricing opacity creates several operational challenges:

• Extended evaluation cycles: Teams cannot self-qualify based on budget before engaging sales
• Budget planning uncertainty: Finance teams lack benchmarks for annual planning
• Competitive comparison difficulty: Side-by-side cost analysis requires multiple vendor negotiations
• Approval process delays: Procurement committees need pricing documentation that may not exist in standard formats

Teams evaluating MCP gateway alternatives should factor this procurement friction into their total timeline and resource planning.

Navigating Quote-Based Services

Questions That May Affect a Custom Quote

Runlayer does not publish a formal rate card or documented formula showing how individual requirements change price. Teams should ask sales how the quote changes based on:

Deployment:

• Managed versus self-hosted deployment
• Private-network or data-residency requirements
• Infrastructure and hosting responsibilities

Scale:

• Number of employees and agents governed
• Connector and MCP requirements
• Usage or capacity limits, if any

Security and Compliance:

• Required security and compliance documentation
• Support and SLA requirements
• Additional implementation or professional services

What to Expect in Enterprise Negotiations

Teams entering custom quote discussions should prepare for:

Information Requests:

• Expected user counts by tier (developers, analysts, executives)
• Projected MCP tool call volumes
• Current AI client deployment (Claude, Cursor, ChatGPT, Copilot usage)
• Infrastructure preferences (managed vs. self-hosted)

Negotiation Variables:

• Contract length (annual vs. multi-year with discounts)
• Payment terms (monthly vs. annual prepay)
• Implementation and onboarding fees
• Professional services for custom integrations

Documentation Gaps:

• Standard pricing sheets may not exist in traditional formats
• ROI calculators may require building internally
• Competitive benchmark data must come from independent sources

For organizations comparing quote-based vendors, Mintive's pricing page provides team-size bands and describes its per-user licensing structure, but final pricing still requires a custom quote.

How to Estimate Runlayer's Total Cost of Ownership

Infrastructure and Operations Overhead

The stated platform fee represents only one component of total cost of ownership. Teams deploying Runlayer must account for:

Self-Hosted Deployment Costs:

• Kubernetes cluster management and scaling
• Database infrastructure (provisioning, backups, monitoring)
• Networking configuration and security groups
• Container orchestration and upgrade management

Connector Runtime Operations:

• Operational responsibility depends on the selected deployment model
• Runlayer can host custom MCP servers on managed infrastructure
• Self-hosted deployments shift more infrastructure, scaling, monitoring, and upgrade responsibility to the customer
• Teams should confirm which connector hosting and OAuth responsibilities are included in their specific quote

Engineering Time Investment:

• Initial deployment timing varies by infrastructure choices
• Ongoing maintenance burden for infrastructure updates
• Integration engineering for custom connectors
• Troubleshooting complexity depends on deployment model

Comparative Deployment Model Analysis

When evaluating total cost of ownership, consider:

Self-Hosted Deployment:

• Platform license: Custom quote
• Infrastructure: Customer-operated infrastructure adds direct cloud and operations costs
• Connector operations: Depends on which connectors and runtimes the customer operates
• Implementation: Depends on identity, policy, connector, and deployment complexity
• Ongoing maintenance: Customer owns more infrastructure maintenance

Managed Deployment:

• Platform license: Custom quote
• Infrastructure: Confirm what infrastructure is included in the vendor quote
• Connector operations: Managed connector hosting can reduce customer runtime operations
• Implementation: Depends on identity, policy, connector, and deployment complexity
• Ongoing maintenance: Vendor owns more of the managed platform operations

No defensible dollar comparison can be calculated from public list prices because both Runlayer and Mintive require custom quotes. Compare actual vendor quotes using the same deployment, user, agent, connector, support, and compliance assumptions.

The True Cost of Configuration Complexity

Beyond direct expenses, teams should evaluate:

• Least-privilege configuration: Time required to set up granular access controls across connector catalogs
• Debugging overhead: Troubleshooting complexity without vendor-managed runtime visibility
• Upgrade management: Self-hosted teams must apply updates manually, adding maintenance windows
• OAuth setup per connector: Authentication configuration requirements in self-hosted environments

Platforms offering hosted MCP connectors can reduce operational burden by managing connector runtime infrastructure.

Runlayer's Governance Approach

Feature Capabilities

Because Runlayer does not publish feature packaging by price tier, buyers should confirm which governance capabilities are included in their specific quote:

Core Platform Features:

• MCP Gateway with an approved connector catalog plus support for internal and custom MCP servers
• Support for major AI clients including Claude Code, Cursor, ChatGPT, VS Code, GitHub Copilot, Codex, and Windsurf
• Identity-aware access policy and audit logging

Security Capabilities:

• ToolGuard scanning around MCP definitions and runtime calls
• AgentGuard session-level monitoring
• Shadow AI discovery on managed devices through MDM deployment

Deployment and Governance Capabilities:

• Managed and self-hosted deployment options
• Identity-aware policy configuration
• Enterprise identity, audit, and deployment controls

How Governance Drives Pricing Decisions

For security-conscious organizations, governance capabilities often influence pricing discussions. Key considerations include:

Compliance Requirements:

• Security attestation documentation availability
• HIPAA compliance documentation and BAA support
• Audit trail completeness for regulatory reporting

Security Feature Value:

• Prompt injection detection and blocking
• PII and secrets scanning on tool calls
• Session monitoring for behavioral anomalies

Access Control Granularity:

• Tool-level permissions per user or agent
• RBAC integration with existing identity providers
• Policy enforcement across multiple AI clients

Teams with stringent security governance requirements should evaluate whether security capabilities align with their compliance mandates before finalizing pricing discussions. The NIST AI Risk Management Framework provides guidance on evaluating AI governance controls.

The Value of Unified Governance

The Virtual MCP Architecture Advantage

An alternative approach to AI agent governance centers on unified primitives that combine multiple governance functions into single manageable objects:

Virtual MCP Governance:

• Virtual MCPs provide per-role or per-use-case endpoints with SCIM-driven membership, curated tools, and access policies
• Agent Bundles extend the model to per-agent identity, scoped tools, M2M authentication, and independent credential rotation or revocation
• Coursera used Virtual MCPs as part of its governed rollout to 2,000 employees in weeks

SCIM-Driven Access Control:

• Directory group membership automatically determines tool access
• User provisioning and deprovisioning propagates through identity provider integration
• Reduces manual access management overhead

Curated Tool Surfaces:

• Each Virtual MCP exposes only relevant tools for specific use cases
• Prevents context-window bloat from exposing unnecessary capabilities
• Read-only vs. read-write access handled through separate bundles over shared connectors

Cost Impact of Unified Architecture

Organizations deploying unified governance primitives report significant efficiency gains:

• Deployment velocity: Teams like Coursera deployed governed AI to 2,000 employees in weeks using Virtual MCP bundles
• Setup time reduction: Virtual MCPs reduce configuration overhead compared to managing separate objects
• Administration simplification: Virtual MCPs centralize role- or use-case-specific tool surfaces, access policies, and directory-driven membership

The MCP Gateway approach of bundling connectors behind governed endpoints can reduce both direct costs and operational complexity.

Per-Agent Identity and Programmable Security

The Agent Identity Advantage

As enterprises scale from 10 to 100+ autonomous agents, identity management becomes a critical governance requirement. Advanced identity features include:

Per-Agent Credentials:

• Dedicated OAuth 2.0 credentials for each autonomous agent
• Independent credential rotation without affecting other agents or users
• Scoped tool access specific to each agent's purpose

Authentication Options:

• Bearer keys for simple static authentication
• M2M tokens via OAuth client-credentials exchange
• Workload identity federation for zero-secret deployments

Audit Attribution:

• Every tool call attributed to the specific agent identity
• Clear separation between human and agent actions in audit logs
• Compliance reporting that distinguishes automated from manual operations

Organizations requiring first-class agent identities should factor identity management capabilities into their evaluation.

Programmable Security

Beyond preset security rules, programmable policy capabilities provide flexibility for unique compliance requirements:

Custom Policy Execution:

• JavaScript middleware running on every tool call
• Ability to inspect, transform, mask, or block requests inline
• Integration with existing DLP investments (AWS Bedrock Guardrails, GCP DLP, Microsoft Purview)

Rule Configuration:

• Pattern matching on tool names, arguments, and content
• Actions including flag, block, ask-user, mask, and notify
• Dry-run modes for policy testing before enforcement

Managed Detection Policies:

• Out-of-box screening for prompt injection, credentials, PII, and harmful content
• Centrally maintained detection models without customer authoring
• Monitoring vs. enforcing mode toggles

Platforms offering programmable guardrails through middleware architectures enable security teams to maintain existing tool investments while extending governance to AI agents. The OWASP AI Security guidance provides frameworks for evaluating AI security controls.

Detecting Shadow AI

The Shadow AI Visibility Problem

One of the most significant considerations in AI governance involves unmanaged AI tool adoption:

What Shadow AI Includes:

• Claude Code, Cursor, or Copilot installations on developer endpoints without IT visibility
• Off-gateway MCP server connections bypassing centralized controls
• Local tool calls (bash commands, file reads) executed by coding agents
• Unauthorized AI tools that employees self-provision

Monitoring Approaches:

Gateway-Only Visibility:

• Sees only traffic routed through the managed gateway
• Misses local activity on developer machines
• Lower implementation complexity

Local Agent Visibility:

• Supported local hooks can capture activity such as commands, file access, prompts, and MCP tool calls
• Supported off-gateway activity can report into centralized monitoring
• Coverage varies by client, agent, and hook phase

Cross-Platform Monitoring:

• Centralized visibility across supported agent environments and governed MCP traffic
• Usage and token-cost tracking where supported
• SIEM export for integration with existing security operations

The Agent Monitor approach extends visibility to capture supported agent activity, with coverage varying by client and hook phase.

Why Teams Choose Mintive for Predictable AI Governance

Organizations evaluating AI governance platforms increasingly prioritize deployment velocity, operational simplicity, and alignment with existing enterprise identity systems. Mintive addresses these requirements through an architecture designed for enterprise rollout at scale.

Deployment Model

Mintive provides fully managed infrastructure that eliminates customer-operated Kubernetes clusters, database provisioning, and connector runtime operations. Virtual MCPs combine SCIM directory integration, curated tool surfaces, and policy enforcement into single governed endpoints, enabling teams like Coursera to deploy AI governance to 2,000 employees in weeks rather than months. Agent Bundles extend this model to autonomous agents with per-agent credentials, independent rotation, and scoped tool access.

Security and Compliance

Mintive is SOC 2 Type II audited, compliant with HIPAA standards, and able to sign BAAs for customers handling PHI. Programmable guardrails support inline JavaScript policy execution with pattern matching, masking, blocking, and integration with existing DLP tools from AWS, GCP, and Microsoft. Supported local hooks capture off-gateway activity including prompts, commands, file access, and MCP tool calls, with coverage varying by client and agent. Audit logs export to Sentinel, Splunk, S3, and other SIEM platforms.

Pricing and Licensing

Mintive publishes team-size bands and uses per-user licensing for active AI-agent users. While dollar pricing requires a custom quote, the per-user model provides predictable cost scaling as deployments grow. Implementation typically involves identity provider integration, Virtual MCP configuration for key roles, policy setup, and client deployment, with documented customer rollouts completing in weeks.

Customer Outcomes

Documented deployments include Stability AI's company-wide AI governance rollout and Modern Treasury's Technical Account Manager productivity gains through governed tool access. Teams report that Virtual MCPs reduce configuration overhead compared to managing separate plugins, access rules, and agent accounts across large connector inventories.

For organizations seeking an alternative to quote-only pricing models, Mintive provides transparent licensing structure with managed infrastructure, unified governance primitives, and enterprise security controls.

Frequently Asked Questions

How does Runlayer's pricing compare to usage-based billing models common in cloud infrastructure?

Runlayer's pricing appears to follow a platform-fee model based on deployment size rather than pure usage-based billing. Teams accustomed to usage-based models should request detailed breakdowns of how deployment size translates to pricing during negotiations, including whether there are overage charges, burst pricing, or volume discounts.

What happens to pricing if an organization needs to rapidly scale AI agent deployment mid-contract?

Contract flexibility for scaling remains unclear given Runlayer's quote-based approach. Organizations planning significant growth should negotiate explicit terms for adding users, agents, or MCP servers mid-contract, including whether additional capacity is available at the contracted rate and what the lead time is for provisioning.

Are there industry-specific pricing considerations for regulated sectors like healthcare or financial services?

Runlayer does not publish whether regulated-industry requirements change pricing. Healthcare and financial-services buyers should ask whether BAA support, audit capabilities, data residency, security documentation, deployment requirements, and vendor-risk review are included in the quoted price or priced separately.

How do connector marketplace and catalog sizes affect overall platform costs?

The practical cost impact depends on how connectors are licensed and operated. Key questions include whether all catalog connectors are included in base pricing, whether there are additional fees for premium connectors, and who operates the connector runtime. Platforms offering hosted connector runtimes bundle operational costs into the platform fee.

What questions should procurement teams ask during vendor pricing discussions to ensure complete cost visibility?

Procurement teams should ask what is included in the platform fee versus add-on pricing, whether hosted connectors are included or the customer must operate runtime infrastructure, professional services fees for deployment, how pricing changes as user counts or tool call volumes increase, whether compliance features are included or premium, minimum commitment lengths and termination provisions, and whether there are overage charges or fees for specific integrations.