specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Protect AI
providerId: protectai
created: '2026-06-20'
modified: '2026-06-20'
reconciled: false
tags:
- AI
- ML
- Security
- LLM
- Guardrails
- FinOps
- Cost Management
- FOCUS
description: >-
  FinOps view of Protect AI spend. Cost falls into two distinct shapes. The
  open-source LLM Guard and ModelScan carry no license or per-call fee; their
  cost is entirely the self-hosted infrastructure (compute, GPU, storage,
  operations) you run them on. The commercial platform (Guardian, Recon, Layer,
  now via Palo Alto Networks Prisma AIRS) is a sales-led subscription whose
  charges are governed by contract and are not publicly metered. Specific
  commercial rates are not reconciled here.
notes: >-
  Track open-source LLM Guard cost as infrastructure spend in your existing
  cloud FinOps practice. Treat commercial Protect AI / Prisma AIRS charges as
  contracted subscription spend; verify meters against your agreement on
  reconciliation.
sources:
- https://github.com/protectai/llm-guard
- https://github.com/protectai/modelscan
- https://protectai.com/
- https://focus.finops.org/focus-specification/v1-3/
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: Protect AI
serviceCategory: Security
billingModel:
  pricingCategory: Subscription
  billingFrequency: Annual
  billingCurrency: USD
  chargeCategories:
  - Usage
  - Purchase
focusColumns:
  ServiceName: Protect AI Platform
  ServiceCategory: Security
  ProviderName: Protect AI
  PublisherName: Protect AI
  InvoiceIssuerName: Palo Alto Networks
  BillingCurrency: USD
  ChargeCategory: Purchase
  PricingCategory: Subscription
meters:
- name: self_hosted_compute_hours
  description: Compute hours running the self-hosted LLM Guard API and ModelScan (your infra cost, not a Protect AI charge).
  unit: hours
  aggregation: sum
  dimensions:
  - account
  - deployment
- name: commercial_subscription
  description: Contracted subscription for Guardian / Recon / Layer; charged per agreement, not publicly metered.
  unit: contract
  aggregation: sum
  dimensions:
  - account
  - product
principles:
- name: Visibility
  description: Attribute self-hosted LLM Guard cost to its host infrastructure tags; track commercial subscription line items from the Protect AI / Palo Alto Networks invoice.
- name: Allocation
  description: Tag LLM Guard deployments per workload/team; map commercial seats and products to internal cost centers.
- name: Optimization
  description: Prefer self-hosted open-source LLM Guard where it meets requirements; right-size scanner sets and worker counts to cut compute; scope commercial products to genuine coverage gaps.
- name: Accountability
  description: Assign owners per deployment and per commercial product; review infra and subscription spend against budget.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com