name: Encord FinOps
description: >
  FinOps Framework 1.0 FOCUS-aligned cost and usage profile for Encord's
  computer vision data labeling and model evaluation platform. Encord uses
  a tiered SaaS pricing model with a free Starter plan, a quote-based Team
  plan, and custom Enterprise agreements. Cost drivers include the number of
  seats, annotation volume, storage consumption, automated labeling agent
  usage, and active learning / model evaluation compute.
specificationVersion: "1.0"
url: https://encord.com/pricing/

provider:
  name: Encord
  website: https://encord.com/
  pricingPage: https://encord.com/pricing/
  billingModel: SaaS subscription with usage-based components
  currency: USD

costDimensions:
  - dimension: Platform Subscription
    description: >
      Base subscription fee per seat or per team, covering access to the
      Encord Annotate and Encord Active modules. Starter tier is free.
      Team and Enterprise pricing is quote-based.
    unit: seat / month
    tiers:
      - tier: Starter
        price: 0
        unit: free
      - tier: Team
        price: null
        unit: quote-based per seat
      - tier: Enterprise
        price: null
        unit: custom contract

  - dimension: Annotation Volume
    description: >
      Cost may scale with the number of annotation tasks, labeled assets,
      or hours of annotator usage depending on the contract structure.
      High-volume annotation pipelines should account for this as a
      primary cost driver.
    unit: labeled asset or annotation hour
    notes: Pricing negotiated per contract; contact Encord sales.

  - dimension: Storage
    description: >
      Encord manages metadata and label storage. Large-scale datasets with
      many storage items (images, video frames, DICOM files, point clouds)
      may incur additional storage costs. Cloud data registered via customer-
      owned AWS S3, Azure Blob, or GCP buckets does not add Encord storage
      costs but may affect cloud provider bills.
    unit: GB / month
    notes: >
      Storage costs for customer data remain with the customer's cloud
      provider. Encord stores labels and metadata internally.

  - dimension: Automated Labeling Agents
    description: >
      Encord supports automated labeling via pre-built and custom agents.
      Enterprise and Team plans may include agent execution costs as part
      of the subscription or charged per inference run depending on the
      agreement.
    unit: agent run
    notes: Included or metered depending on plan tier.

  - dimension: Model Evaluation Compute (Encord Active)
    description: >
      Encord Active computes embeddings, quality metrics, and similarity
      indexes for datasets. Compute costs for large datasets with many
      assets may be a factor in Enterprise agreements.
    unit: compute hours or dataset size
    notes: Included in Team and Enterprise plans; quotas may apply.

finOpsGuidance:
  - recommendation: >
      Start with the free Starter plan to evaluate Encord's capabilities
      before committing to a Team or Enterprise contract.
  - recommendation: >
      For cost predictability, negotiate a fixed-seat Enterprise agreement
      rather than usage-based billing if annotation volume is high and
      consistent.
  - recommendation: >
      Register data using customer-owned cloud storage (S3, Azure, GCP)
      to avoid vendor lock-in and control storage costs independently of
      the Encord subscription.
  - recommendation: >
      Monitor automated labeling agent usage; agents can accelerate
      throughput but may increase metered costs if not capped by contract.
  - recommendation: >
      Use the Python SDK's bulk operations (get_storage_items,
      find_storage_items) to minimize API call overhead and associated
      compute costs when processing large datasets.