# Vendor facets — Mintlify. A hosted docs platform that now ships most of the agent-discovery
# surface by default: llms.txt, a generated skill.md, an A2A agent card, a docs MCP server and an
# RFC 9727 api-catalog on every public site. The lift lands in Agent Readiness and a few
# discoverability / ergonomics pointers; it cannot write the OpenAPI, so Contract Quality (the
# heaviest facet) is untouched, and the MCP server is one template graded `platform`.
vendor: mintlify
name: Mintlify
website: https://mintlify.com
areas:
- documentation
- mcp-hosting
registry_keys:
- mintlify
- mintlify-mcp
rubric_schema_version: 0.22.0
generated: '2026-09-25'
features_refreshed: '2026-09-25'
basis: measured
summary: >-
  Mintlify's lift is concentrated in the standalone Agent Readiness layer: a generated skill.md, a conformant
  A2A agent card and a docs MCP server on every public site, plus llms.txt and an RFC 9727 api-catalog
  that also move Discoverability. The biggest qualifier is the MCP server: it is one Mintlify template
  serving docs search, so the rubric grades it `platform` (a quarter credit), and none of it touches Contract
  Quality, because Mintlify renders the OpenAPI a provider brings and never writes it. Most of the credit
  also depends on serving the docs from the provider's own custom domain rather than a mintlify.app subdomain.
features:
- id: llms-txt
  name: Auto-generated llms.txt and llms-full.txt
  description: >-
    Hosts an always-current llms.txt index and llms-full.txt corpus at the docs root (and /.well-known/llms.txt),
    overridable with a custom file.
  source: https://www.mintlify.com/docs/ai/llmstxt
  tier: all
- id: skill-md
  name: Generated skill.md and agent-skills discovery index
  description: >-
    An agentic loop writes a skill.md from the provider's docs and publishes it at /skill.md and /.well-known/agent-skills/;
    a provider-written skill.md overrides it.
  source: https://www.mintlify.com/docs/ai/skillmd
  tier: all
- id: agent-card
  name: A2A agent card per docs site
  description: Serves /.well-known/agent-card.json (A2A 0.3 shape) listing the site's skills.
  source: https://www.mintlify.com/docs/ai/skillmd
  tier: all
- id: search-mcp
  name: Hosted search MCP server
  description: >-
    Every site gets a Mintlify-hosted MCP server at <docs-host>/mcp with search, docs-filesystem query
    and submit-feedback tools, advertised through /.well-known/mcp and server-card endpoints.
  source: https://www.mintlify.com/docs/ai/model-context-protocol
  tier: all
- id: api-catalog
  name: RFC 9727 api-catalog
  description: >-
    Serves /.well-known/api-catalog as a linkset of the OpenAPI documents ingested from docs.json, advertised
    in a Link header on every response.
  source: https://www.mintlify.com/docs/help-center/register-external-mcp-server-in-discovery
  tier: all
- id: markdown-export
  name: Markdown twins and Accept negotiation
  description: >-
    Every page is served as Markdown at <url>.md and to Accept text/markdown requests, with a Markdown
    404 that points agents back to llms.txt.
  source: https://www.mintlify.com/docs/ai/markdown-export
  tier: all
- id: webmcp
  name: WebMCP tool registration
  description: Registers the docs tools on the page through navigator.modelContext for in-browser agents.
  source: https://www.mintlify.com/docs/ai-native
  tier: paid
- id: changelog
  name: Changelog pages with RSS
  description: Update components build a dated changelog page with tag filters and an RSS feed.
  source: https://www.mintlify.com/docs/create/changelogs
  tier: all
- id: api-playground
  name: Interactive API playground
  description: Generated OpenAPI/AsyncAPI reference pages with a request builder that sends real calls.
  source: https://www.mintlify.com/docs/api-playground/overview
  tier: all
- id: sdk-examples
  name: SDK code samples in the reference
  description: Renders x-codeSamples from the OpenAPI (or Speakeasy-generated samples) beside each endpoint.
  source: https://www.mintlify.com/docs/api-playground/adding-sdk-examples
  tier: all
- id: hosted-docs
  name: Hosted documentation site and API reference
  description: The docs site itself, with guides and an OpenAPI-generated API reference on a custom domain.
  source: https://www.mintlify.com/docs/api-playground/openapi-setup
  tier: all
- id: page-feedback
  name: Page feedback from readers and agents
  description: Feedback widgets, ratings and agent-submitted feedback routed to the dashboard.
  source: https://www.mintlify.com/docs/optimize/feedback
  tier: all
- id: visibility-component
  name: Human- vs agent-only content
  description: >-
    A Visibility component that shows content only on the web page or only in the Markdown served to agents.
  source: https://www.mintlify.com/docs/components/visibility
  tier: all
- id: x-mint
  name: x-mint OpenAPI extension
  description: >-
    An OpenAPI extension that ties operations and properties to docs navigation and user-group visibility.
  source: https://www.mintlify.com/docs/api-playground/openapi-setup
  tier: all
- id: auth-prm
  name: OAuth protected-resource metadata for authenticated docs
  description: Sites behind docs authentication advertise an oauth-protected-resource document in the
    Link header.
  source: https://www.mintlify.com/docs/ai/llmstxt
  tier: unknown
maps:
- feature: llms-txt
  check: llms_txt_published
  layer: composite
  credit: 1.0
  provider_must: >-
    Serve the docs from its own custom domain; on a bare mintlify.app subdomain the file is not on the
    provider's surface.
  note: >-
    Generated, but generated from the provider's own documentation and served on the provider's domain,
    so it reads as the provider's llms.txt (rule 2's own-content, own-domain case).
  catalog_pass_rate: 0.351
  facet: discoverability
  points: 4
  baseline_pass_rate: 0.65
  cohort_pct: 79.5
  control_pct: 64.7
  delta_pp: 14.8
- feature: skill-md
  check: agent_skills
  layer: agent_readiness
  credit: 1.0
  provider_must: >-
    Keep the docs public and on its own custom domain (the skill is only generated for public sites);
    writing its own skill.md is better still.
  note: >-
    Authored by Mintlify's model from the provider's docs and served on the provider's domain, so first-party
    under rule 2. On a mintlify.app host the same file grades `derived` (0.25); the 2026-09-14 cohort
    split 59 first-party to 32 derived for exactly that reason.
  points: 5
  baseline_pass_rate: 0.236
  cohort_pct: 37.3
  control_pct: 23.3
  delta_pp: 14.0
- feature: skill-md
  check: agent_skill_present
  layer: composite
  credit: 1.0
  provider_must: >-
    Declare the skill URL as an AgentSkill pointer in apis.yml; same custom-domain caveat as agent_skills.
  note: The same artifact scored a second time; do not count the work behind it twice.
  catalog_pass_rate: 0.033
  facet: developer_ergonomics
  points: 3
  baseline_pass_rate: 0.236
  cohort_pct: 37.3
  control_pct: 23.3
  delta_pp: 14.0
- feature: search-mcp
  check: mcp_server
  layer: agent_readiness
  grade: platform
  note: >-
    One Mintlify template across the customer base, answering docs search rather than calling the API,
    so 0.22.0 grades it `platform` (0.25 of 12). The case against the discount is that it runs on the
    customer's domain from the customer's content; that is a rubric argument, not a credit this file can
    take.
  points: 12
  baseline_pass_rate: 0.22
  cohort_pct: 45.0
  control_pct: 21.5
  delta_pp: 23.5
- feature: agent-card
  check: agent_card
  layer: agent_readiness
  grade: conformant
  provider_must: Serve the docs at a host root (a reverse proxy must forward /.well-known/agent-card.json).
  note: >-
    Carries capabilities, protocolVersion, skills and the transport/mode fields, so the harvest grades
    it conformant; in the catalog 28 of 36 Mintlify-referencing cards graded conformant, the rest near-conformant
    or flavored. It describes the docs, not the API.
  points: 8
  baseline_pass_rate: 0.03
  cohort_pct: 3.3
  control_pct: 3.0
  delta_pp: 0.3
- feature: api-catalog
  check: well_known_published
  layer: composite
  credit: 1.0
  provider_must: >-
    Serve the docs at the root of a host the harvest probes (docs.<domain>); under a /docs base path the
    catalog sits at /docs/.well-known/api-catalog, which a host-root probe never reads.
  note: >-
    A real RFC 9727 linkset, but its service-desc links are presigned S3 URLs that expire after an hour,
    so the catalog points at contracts an agent cannot fetch later. In a 32-provider sample of Mintlify
    MCP customers, 8 had a 200 api-catalog recorded.
  catalog_pass_rate: 0.005
  facet: discoverability
  points: 6
  baseline_pass_rate: 0.018
  cohort_pct: 0.9
  control_pct: 1.8
  delta_pp: -0.9
- feature: api-catalog
  check: well_known_catalog
  layer: agent_readiness
  provider_must: Declare the api-catalog URL as an APICatalog / WellKnown pointer in apis.yml.
  points: 4
  baseline_pass_rate: 0.061
  cohort_pct: 3.6
  control_pct: 6.2
  delta_pp: -2.6
- feature: changelog
  check: change_log_present
  layer: composite
  provider_must: Actually keep the changelog page and declare it as a ChangeLog pointer.
  catalog_pass_rate: 0.157
  facet: operational_transparency
  points: 6
  baseline_pass_rate: 0.438
  cohort_pct: 50.9
  control_pct: 43.7
  delta_pp: 7.2
- feature: api-playground
  check: console_or_sandbox
  layer: composite
  provider_must: >-
    Declare the playground as a Console / Playground pointer; it is only as live as the servers in the
    provider's OpenAPI.
  catalog_pass_rate: 0.089
  facet: developer_ergonomics
  points: 3
  baseline_pass_rate: 0.332
  cohort_pct: 28.6
  control_pct: 33.3
  delta_pp: -4.7
- feature: sdk-examples
  check: openapi_examples
  layer: agent_readiness
  grade: documented
  partial: true
  partial_note: >-
    The dimension reads examples inside the OpenAPI first. Mintlify renders what the spec holds, so on
    its own it reaches only the Examples / Code Samples link fallback.
  points: 7
  baseline_pass_rate: 0.36
  cohort_pct: 44.4
  control_pct: 35.9
  delta_pp: 8.5
- feature: hosted-docs
  check: documentation_present
  layer: composite
  provider_must: Declare the docs root as a Documentation pointer.
  catalog_pass_rate: 0.453
  facet: developer_ergonomics
  points: 4
  baseline_pass_rate: 0.95
  saturated: true
  saturated_note: >-
    95% of providers with a contract, docs and a reference already earn this; the vendor cannot move it
    for most of its buyers.
  cohort_pct: 100.0
  control_pct: 94.9
  delta_pp: 5.1
- feature: hosted-docs
  check: api_reference_present
  layer: composite
  provider_must: Declare the generated reference as an APIReference pointer.
  catalog_pass_rate: 0.222
  facet: developer_ergonomics
  points: 3
  baseline_pass_rate: 0.942
  saturated: true
  saturated_note: >-
    94% of providers with a contract, docs and a reference already earn this; the vendor cannot move it
    for most of its buyers.
  cohort_pct: 100.0
  control_pct: 94.0
  delta_pp: 6.0
- feature: hosted-docs
  check: portal_present
  layer: composite
  provider_must: >-
    Declare the site as a DeveloperPortal pointer; a docs site alone is a portal only once the provider
    says so.
  catalog_pass_rate: 0.228
  facet: developer_ergonomics
  points: 4
  baseline_pass_rate: 0.633
  cohort_pct: 75.4
  control_pct: 63.0
  delta_pp: 12.4
earns_nothing:
- feature: sdk-examples
  check: examples_present
  why: >-
    Contract-side examples are counted in the OpenAPI. Mintlify renders the examples a spec already has
    and writes none back; the +10.9pp measured in 2026-09 was the cohort's own specs.
- feature: search-mcp
  check: well_known_published
  why: >-
    The /.well-known/mcp and server-card documents are none of the four files the check reads (api-catalog,
    security.txt, protected-resource, AAuth). Only the api-catalog row earns it.
- feature: page-feedback
  check: support_channel
  why: A feedback widget is not a Support, HelpCenter or Community channel.
- feature: auth-prm
  check: protected_resource_metadata
  layer: agent_readiness
  why: >-
    It describes the documentation site's login, not the API's resource server. If a probe credits it,
    that is a false credit the provider did not earn for its API.
out_of_reach:
  checks:
  - contract_present
  - info_complete
  - operations_description_coverage
  - error_responses_documented
  - sdk_count_1
  - sdk_count_3
  - cli_present
  - postman_present
  - status_page_present
  - rules_present
  note: >-
    Mintlify renders a contract and hosts prose; it does not author OpenAPI, generate SDKs or a CLI, publish
    a Postman collection, run a status page or ship a ruleset.
unscored_practice:
- feature: x-mint
  why: Contract-to-docs pointer inside the OpenAPI; 46 providers ship it and no check reads it.
- feature: markdown-export
  why: >-
    Markdown twins and Accept text/markdown negotiation; the rubric notes markdown negotiation and scores
    it nowhere.
- feature: webmcp
  why: In-browser tool registration via navigator.modelContext; no check reads WebMCP.
- feature: visibility-component
  why: Agent-only content blocks are real agent-facing authoring with no scored signal.
- feature: llms-txt
  why: llms-full.txt and the Link / X-Llms-Txt discovery headers earn nothing beyond the llms.txt row.
surface:
  discoverability:
    reachable: 10.0
    total: 54
  operational_transparency:
    reachable: 6.0
    total: 38
  developer_ergonomics:
    reachable: 17.0
    total: 42
  agent_readiness:
    reachable: 21.8
    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://www.mintlify.com/docs/ai-native
- https://www.mintlify.com/docs/ai/llmstxt
- https://www.mintlify.com/docs/ai/markdown-export
- https://www.mintlify.com/docs/ai/model-context-protocol
- https://www.mintlify.com/docs/ai/skillmd
- https://www.mintlify.com/docs/api-playground/adding-sdk-examples
- https://www.mintlify.com/docs/api-playground/openapi-setup
- https://www.mintlify.com/docs/api-playground/overview
- https://www.mintlify.com/docs/components/visibility
- https://www.mintlify.com/docs/create/changelogs
- https://www.mintlify.com/docs/help-center/register-external-mcp-server-in-discovery
- https://www.mintlify.com/docs/optimize/feedback
measured:
  cohort:
    method: vendors-catalog.json detections (CNAME / header / URL shape / markup), never a name match
    detected: 195
    in_baseline: 112
  control:
    basis: providers earning contract_present + documentation_present + api_reference_present, minus the
      cohort
    n: 5104
  metric: >-
    cohort_pct / control_pct = mean share of the check's points earned (derived and platform credit weighted),
    x100
  measured_on: '2026-09-25'
  composite_mean:
    cohort: 49.4
    control: 48.0
  agent_readiness_mean:
    cohort: 37.2
    control: 32.4
  status: measured
  caveat: >-
    A cohort delta is association, not cause: customers choose a vendor for reasons that also move their
    score. Read it beside the capability map, never instead of it.
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: 8975
  conditional_rows: excluded (they depend on what the API already does)
  composite_lift:
    median: 7.2
    p75: 8.7
    p90: 10.1
    max: 12.1
    mean_among_movers: 6.9
  agent_readiness_lift:
    median: 14.4
    p75: 15.6
    p90: 15.7
    max: 18.2
    mean_among_movers: 14.0
  facet_lift_median_among_movers:
    discoverability: 11.1
    operational_transparency: 15.8
    developer_ergonomics: 21.5
  composite_band_moves:
    thin -> developing: 2398
    developing -> strong: 1267
    emerging -> thin: 409
    strong -> exemplar: 319
    minimal -> emerging: 3
  agent_readiness_band_moves:
    agent-aware -> agent-ready: 4914
    agent-ready -> agent-native: 316
    agent-aware -> agent-native: 72
  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