specification: FinOps Framework specificationVersion: '1.0' schema: https://www.finops.org/framework/ provider: Barndoor providerId: barndoor created: '2026-05-15' modified: '2026-05-15' reconciled: false tags: - AI Agents - AI Governance - MCP - Control Plane - FinOps - FOCUS description: >- FOCUS 1.3-aligned FinOps for Barndoor. Public pricing is seat-style (non- human identities) and feature-gated by plan tier, not consumption-metered. The pricing page shows $500/month Team, custom Pro and Enterprise, and free Trial. The likely commercial meters are: registered agents (seats), governed MCP server registrations, OAuth connection events, governance / policy decisions, and proxied MCP / SSE requests, but only the agent seat ceiling is published. Cost is currently fixed-per-plan; reconciled:false until a real customer contract confirms add-on or overage line items. sources: - https://barndoor.ai/pricing - https://docs.barndoor.ai/api-reference/introduction - https://docs.barndoor.ai/how-tos/log-export alignedWith: framework: FinOps Foundation Framework frameworkUrl: https://www.finops.org/framework/ dataSpec: FOCUS dataSpecVersion: '1.3' dataSpecUrl: https://focus.finops.org/focus-specification/v1-3/ publisherName: Barndoor AI, Inc. serviceCategory: AI Governance / Control Plane billingModel: pricingCategory: Subscription (Plan-Based Seats + Feature Tiers) billingFrequency: Monthly (Annual Commit Available) billingCurrency: USD chargeCategories: - Purchase - Usage - Adjustment - Tax - Credit focusColumns: ServiceName: Barndoor Platform ServiceCategory: AI Governance / Control Plane ProviderName: Barndoor PublisherName: Barndoor AI, Inc. InvoiceIssuerName: Barndoor AI, Inc. BillingCurrency: USD ChargeCategory: Purchase meters: - name: plan_subscription description: >- Recurring monthly subscription per plan tier (Trial / Team / Pro / Enterprise). Team is published at $500/month billed annually; Pro and Enterprise are custom. unit: month aggregation: sum dimensions: - organization_id - plan_tier - name: non_human_identities description: >- Registered agents (non-human identities). The published structural quota: Trial unlimited, Team 250, Pro 1,000, Enterprise custom. Likely acts as a seat license rather than a per-call meter. unit: agent aggregation: max dimensions: - organization_id - agent_type - name: mcp_server_registrations description: >- Registered MCP server instances under the organization. Not separately priced today; included in the plan. unit: mcp_server aggregation: max dimensions: - organization_id - server_type - name: oauth_connections description: >- OAuth connections established from agents to backend SaaS through Barndoor's connection broker. Not separately priced today. unit: connection aggregation: sum dimensions: - organization_id - mcp_server_id - status - name: governance_events description: >- Policy decisions emitted by the runtime enforcement engine (Cerbos PDP) against agent actions. Surfaced through audit logs and dashboards; not separately priced today. unit: decision aggregation: sum dimensions: - organization_id - policy_id - authorized - name: mcp_proxy_requests description: >- Requests proxied through Barndoor's MCP / SSE endpoints (`/mcp/{server}`, `/sse/{server}`) to backend MCP servers. Not separately priced today. unit: request aggregation: sum dimensions: - organization_id - mcp_server_id - agent_id - name: audit_log_events_exported description: >- Audit events exported to S3-compatible storage. Batched every 30 seconds or per 100 events. Not separately priced today; included in Pro and Enterprise. unit: event aggregation: sum dimensions: - organization_id - bucket - name: private_deployment description: >- Private cloud / on-premises deployment add-on (Enterprise only). Custom-quoted. unit: month aggregation: max dimensions: - organization_id principles: - name: Visibility description: >- Use Barndoor's audit dashboards plus the audit-log export to S3 / Datadog-compatible JSON to reconcile policy decisions and proxied MCP requests against agent attribution. Pull `getAgentCounts` for current seat utilization. - name: Allocation description: >- Tag agents by team / product via `application_directory_id` so MCP proxy and policy-decision volume flows back to the right cost center. Maintain separate test and production organizations to keep development activity off the production seat count. - name: Optimization description: >- Right-size the plan tier to actual agent count (Trial -> Team -> Pro -> Enterprise). Archive or unregister unused agents to stay within the tier ceiling. Use ABAC conditions to scope tool exposure narrowly so agents only consume the MCP tools they need. - name: Accountability description: >- Assign an AI Governance owner per organization. Review SOC 2 evidence and audit-log export config quarterly; track plan-tier renewals against the agent seat trend. maintainers: - FN: Kin Lane email: kin@apievangelist.com