# Vendor facets — Fastly. CDN plus bot management with an AI-bot monetize option, API schema
# enforcement, semantic caching for LLM APIs, and an open-source Compute recipe that serves
# Markdown to agents. It enforces and caches but never publishes a scored artifact for the
# customer: no maps. Fastly's own MCP server and agent skills score Fastly, not its customers.
vendor: fastly
name: Fastly
website: https://www.fastly.com
areas:
- cdn
registry_keys:
- fastly
rubric_schema_version: 0.22.0
generated: '2026-09-25'
features_refreshed: '2026-09-25'
basis: capability
summary: >-
  Fastly moves no Kin Score check for a customer. Bot management can allow, rate-limit, block or monetize
  AI traffic, and API Enforcement validates requests against a schema the customer supplies. Both are
  enforcement. The markdown-for-agents example is an open-source Compute service the customer deploys
  itself, not a product feature, and 0.22.0 does not score markdown negotiation anyway. Fastly's MCP server
  and Agent Toolkit are for managing Fastly, so they count toward Fastly's own agent readiness, not its
  customers'.
features:
- id: bot-management
  name: Bot Management and AI Bot Management
  description: >-
    Edge detection of AI crawlers, fetchers and headless browsers, answered with dynamic challenges, deception,
    rate limits, blocks or AI monetization.
  source: https://www.fastly.com/products/bot-management
  tier: paid
- id: api-enforcement
  name: API Enforcement
  description: >-
    Validates incoming API requests against schemas the customer defines, and logs or blocks those that
    do not conform.
  source: https://www.fastly.com/products/ai
  tier: paid
- id: ai-runtime-control
  name: AI Runtime Control
  description: A control plane for AI traffic with budget caps, rate limits and per-session attribution.
  source: https://www.fastly.com/products/ai
  tier: paid
- id: ai-accelerator
  name: AI Accelerator
  description: Semantic caching for LLM APIs at the edge.
  source: https://docs.fastly.com/products/ai-accelerator
  tier: paid
- id: fastly-mcp-and-toolkit
  name: Fastly MCP server and Agent Toolkit
  description: An MCP server and open-source agent skills for managing Fastly infrastructure in natural
    language.
  source: https://www.fastly.com/products/ai
  tier: all
- id: html-to-markdown-recipe
  name: html-2-md Compute recipe
  description: >-
    An open-source Compute service, deployed by the customer, that serves Markdown to AI agents in front
    of an HTML origin.
  source: https://www.fastly.com/blog/give-ai-agents-markdown-they-actually-want
  tier: open-source
maps: []
earns_nothing:
- feature: bot-management
  check: agentic_commerce
  layer: agent_readiness
  why: >-
    Monetizing AI traffic at the edge publishes no /.well-known/ucp.json or acp.json, and the fetched
    page names no protocol.
- feature: api-enforcement
  check: contract_present
  layer: composite
  why: It consumes a schema the customer supplies and publishes none.
- feature: ai-runtime-control
  check: rate_limit_signal
  layer: agent_readiness
  why: >-
    Enforced rate limits are not the RateLimit headers documented in the provider's OpenAPI that the dimension
    reads.
- feature: fastly-mcp-and-toolkit
  check: mcp_server
  layer: agent_readiness
  why: >-
    This MCP server exposes Fastly's own API. It counts toward Fastly's agent readiness, not toward a
    customer's.
out_of_reach:
  checks:
  - consent_identity
  - agentic_commerce
  - well_known_published
  - llms_txt_published
  - mcp_server
  - protected_resource_metadata
  - contract_present
  note: >-
    Fastly ships no managed robots.txt, security.txt, llms.txt, OAuth discovery or customer MCP hosting
    on the fetched pages.
unscored_practice:
- feature: html-to-markdown-recipe
  why: >-
    Markdown negotiation for agents is unscored at 0.22.0. This is also a recipe the customer runs, not
    a Fastly product setting.
surface:
  agent_readiness:
    reachable: 0
    total: 139
hard_rule: >-
  A model, not a score. Adopting this vendor changes a provider's Kin Score only when the provider publishes
  the resulting artifacts on its own surface; nothing here writes a score, and no sponsorship or partnership
  can.
method: searched
source:
- https://docs.fastly.com/products/ai-accelerator
- https://www.fastly.com/blog/give-ai-agents-markdown-they-actually-want
- https://www.fastly.com/products/ai
- https://www.fastly.com/products/bot-management
measured:
  cohort:
    method: vendors-catalog.json detections (CNAME / header / URL shape / markup), never a name match
    detected: 41
    in_baseline: 10
  control:
    basis: providers earning contract_present + documentation_present + api_reference_present, minus the
      cohort
    n: 5206
  metric: >-
    cohort_pct / control_pct = mean share of the check's points earned (derived and platform credit weighted),
    x100
  measured_on: '2026-09-25'
  status: 'not measurable: 10 detected customers clear the baseline (need 20)'
simulation:
  simulated_on: '2026-09-25'
  rubric: 0.23.0
  population: providers publishing a contract (contract_present earned), replayable exactly
  providers: 8977
  providers_unreplayable: 987
  providers_moved: 0
  conditional_rows: excluded (they depend on what the API already does)
  composite_lift:
    median: 0.0
    p75: 0.0
    p90: 0.0
    max: 0.0
    mean_among_movers: 0.0
  agent_readiness_lift:
    median: 0.0
    p75: 0.0
    p90: 0.0
    max: 0.0
    mean_among_movers: 0.0
  facet_lift_median_among_movers: {}
  composite_band_moves: {}
  agent_readiness_band_moves: {}
  method: >-
    each provider's own kin/checks file, the vendor's maps at their stated credit, the scorer's composite
    formula; from -> to, nothing written