name: Labelbox FinOps
description: >
  FinOps framework for managing Labelbox API and platform costs.
  Labelbox uses a consumption-based pricing model denominated in Labelbox
  Units (LBUs). Understanding and forecasting LBU consumption is the primary
  cost-management lever for teams using Labelbox at scale.
specificationVersion: "1.0"
framework: FinOps Foundation FOCUS-aligned

costModel:
  type: consumption-based
  currency: USD
  unit: LBU (Labelbox Unit)
  notes: >
    The Free tier provides access to core features at no cost with user,
    project, and ontology limits. The Subscription tier is priced per LBU
    consumed; enterprise volume discounts are available through the sales team.
  pricingURL: https://labelbox.com/pricing/

tiers:
  - name: Free
    monthlyCost: 0
    currency: USD
    description: No-cost tier for evaluation, education, and small teams.

  - name: Subscription
    monthlyCost: variable
    currency: USD
    pricingModel: per-LBU
    description: >
      Consumption-based billing using Labelbox Units. Contact Labelbox sales
      for current per-LBU rates and volume discounts.
    contactURL: https://labelbox.com/contact/

costDrivers:
  - name: Data Labeling Volume
    description: Number of assets labeled and annotations created.
    unit: labeled assets / annotations
  - name: Model-Assisted Labeling
    description: AI-assisted label suggestions consumed.
    unit: AI-assisted labels
  - name: Auto-labeling
    description: Fully automated labeling operations (Subscription only).
    unit: automated labels
  - name: Foundry Model Usage
    description: Inference calls to Foundry models for RLHF and preference data (Subscription only).
    unit: model inference calls
  - name: Storage
    description: Dataset and annotation storage within the Labelbox platform.
    unit: GB stored
  - name: Users
    description: Free tier capped at 30; Subscription provides unlimited users.
    unit: active users

costOptimization:
  - recommendation: Use the Free tier for prototyping and evaluation before committing to Subscription.
  - recommendation: Use bulk operations via the Python SDK to minimize API overhead.
  - recommendation: Leverage model-assisted labeling to reduce human annotation volume and associated costs.
  - recommendation: Use data curation tools to prioritize high-value assets and avoid labeling redundant data.
  - recommendation: Set up Labelbox Monitor (Subscription) for proactive cost and quality alerts.
  - recommendation: Negotiate volume discounts with Labelbox sales for enterprise workloads.

references:
  - url: https://labelbox.com/pricing/
    description: Official Labelbox pricing page
  - url: https://labelbox.com/contact/
    description: Contact Labelbox sales for enterprise pricing