vendors.yaml

# SaaS webhook event catalog — v1 vendor list (30 vendors)
#
# Tier 1 (10): bespoke manual extractors. Vendor has stable, structured docs (single canonical
# page or well-formed reference site) where a targeted scraper + parser can reliably extract
# every event + its payload schema with high confidence.
#
# Tier 2 (20): LLM-assisted extraction. Docs are unstructured prose, scattered across multiple
# pages, or rendered client-side. The pipeline fetches each docs page, feeds it to an LLM with
# a strict JSON output schema, and pins the result with a confidence score that gates inclusion.
#
# selection criteria for v1:
#   - team-ops category (collaboration, dev tools, payments, support, ATS, CRM, HRIS)
#   - public webhook docs (no NDA / paywall)
#   - non-trivial event count (≥5 distinct event types) so the dataset stays useful per-vendor
#   - active product (signals: recent docs updates, active changelog within last 6 months)
#
# fields:
#   slug             — kebab-case identifier, used in URLs and as the row's `vendor` field
#   display_name     — human-readable name
#   category         — high-level grouping for the companion site index page facets
#   tier             — 1 (manual extractor) | 2 (llm-assisted)
#   docs_url         — canonical docs entry point. extractors start here.
#   notes            — scope warnings, scraping strategy flags, decisions deferred to sub 1.5
#
# decisions deferred to sub 1.5 (vendor-scope) are flagged in `notes` with `DECISION:` prefix.

tier_1:
  - slug: slack
    display_name: Slack
    category: collaboration
    tier: 1
    docs_url: https://docs.slack.dev/reference/events
    notes: |
      Events API is the canonical webhook surface. Each event has a dedicated docs page under
      /events/<event_name>/ with structured payload schema in a sidebar. Scraper iterates the
      index page, follows per-event links, parses the schema block.
      DECISION (sub 1.5): include only Events API. Exclude legacy outgoing webhooks (deprecated)
      and Slash Commands (request/response, not event-driven).

  - slug: github
    display_name: GitHub
    category: dev-tools
    tier: 1
    docs_url: https://docs.github.com/en/webhooks/webhook-events-and-payloads
    notes: |
      Single canonical page listing every webhook event with full payload examples and field
      tables. HTML structure is consistent (each event is an h2 with field tables underneath).
      Custom HTML parser. High extraction confidence expected.
      DECISION (sub 1.5): include core webhooks (repository/org/app). Exclude GitHub Actions
      workflow_run details (sub-events that overlap and bloat the catalog).

  - slug: stripe
    display_name: Stripe
    category: payments
    tier: 1
    docs_url: https://docs.stripe.com/api/events/types
    notes: |
      Largest single-vendor event surface in the catalog (~250 event types). Single canonical
      page lists every event_type with one-line trigger description. Payload schemas live on
      separate `Event object` and resource-object pages and must be joined by resource name.
      DECISION (sub 1.5): include all event_types. Schema field uses the resource object schema
      from /api/<resource> docs, not example payloads. Skip Stripe Connect-only events from v1
      and flag for v2 (separate noisy surface).

  - slug: hubspot
    display_name: HubSpot
    category: crm
    tier: 1
    docs_url: https://developers.hubspot.com/docs/api/webhooks
    notes: |
      Webhooks subscribe to object events (contact.creation, deal.propertyChange, etc.). Event
      list is enumerable from a single subscriptions reference page; payload schema is uniform
      across event types (same envelope) so schema extraction is one-shot, then per-event rows
      are generated by enumerating object types × event triggers.
      DECISION (sub 1.5): include CRM object webhooks. Exclude CMS webhooks (separate product,
      separate docs surface, low overlap with team-ops audience).

  - slug: linear
    display_name: Linear
    category: dev-tools
    tier: 1
    docs_url: https://linear.app/developers/webhooks
    notes: |
      Webhook docs are tight and well-structured. Event types are enumerable (Issue, Comment,
      Project, Cycle, Reaction, ProjectUpdate). Payload schema follows Linear's GraphQL types
      which are introspectable. Extractor can join docs prose with GraphQL schema introspection
      for high-confidence rows.

  - slug: notion
    display_name: Notion
    category: collaboration
    tier: 1
    docs_url: https://developers.notion.com/reference/webhooks
    notes: |
      Webhooks reference page enumerates event types with payload examples. Structured docs.
      Smaller surface (~15 event types). Manual extractor straightforward.

  - slug: calendly
    display_name: Calendly
    category: scheduling
    tier: 1
    docs_url: https://developer.calendly.com/api-docs/d7755e2f9e5fe-webhook-subscriptions
    notes: |
      Webhook events documented in API reference. Small but high-intent surface (invitee.created,
      invitee.canceled, etc.). Schema extracted from event object reference.

  - slug: intercom
    display_name: Intercom
    category: support
    tier: 1
    docs_url: https://developers.intercom.com/docs/references/webhooks/webhook-models/
    notes: |
      Webhook models page lists each topic with example payload + schema notes. Topic naming is
      consistent (conversation.user.created, contact.tag.created, etc.) and enumerable from a
      single docs page.

  - slug: zendesk
    display_name: Zendesk
    category: support
    tier: 1
    docs_url: https://developer.zendesk.com/api-reference/webhooks/event-types/event-type-list/
    notes: |
      Event types list page is the canonical surface. Each event type links to a per-event
      payload schema page. Two-step crawl: enumerate index, fetch each child.

  - slug: salesforce
    display_name: Salesforce
    category: crm
    tier: 1
    docs_url: https://developer.salesforce.com/docs/atlas.en-us.api_streaming.meta/api_streaming/intro_stream.htm
    notes: |
      Salesforce uses Platform Events + Change Data Capture + Outbound Messages; there is no
      single "webhooks" surface like Stripe or Slack. Each of those mechanisms has its own docs
      tree.
      DECISION (sub 1.5): include Platform Events catalog (standard events shipped by SF) and
      Change Data Capture standard objects. Exclude Outbound Messages (SOAP, legacy, low intent)
      and per-org custom Platform Events (not enumerable from public docs). Tier this vendor at
      1 because the standard-event surface is enumerable from a single docs page; custom events
      are explicitly out of v1 scope.

tier_2:
  - slug: asana
    display_name: Asana
    category: collaboration
    tier: 2
    docs_url: https://developers.asana.com/docs/webhooks
    notes: |
      Webhook events documented in prose. Payload structure documented via generic "resource"
      shape. LLM extraction with confidence gate.

  - slug: jira
    display_name: Jira (Cloud)
    category: dev-tools
    tier: 2
    docs_url: https://developer.atlassian.com/cloud/jira/platform/webhooks/
    notes: |
      Atlassian docs are sprawling. Event list is in a sidebar; payload schemas are scattered
      across multiple linked pages and overlap with Jira REST API resource docs.
      DECISION (sub 1.5): Jira Cloud only. Server/Data Center webhooks are a separate product
      with different event surface — exclude from v1.

  - slug: microsoft-teams
    display_name: Microsoft Teams
    category: collaboration
    tier: 2
    docs_url: https://learn.microsoft.com/en-us/microsoftteams/platform/webhooks-and-connectors/what-are-webhooks-and-connectors
    notes: |
      MS Learn docs are deeply nested. Outgoing webhooks, incoming webhooks, and Graph
      change notifications are three separate mechanisms.
      DECISION (sub 1.5): include Graph change notifications for Teams resources only (chats,
      messages, channels). Exclude outgoing/incoming webhook templates (those are user-defined,
      not platform events).

  - slug: discord
    display_name: Discord
    category: collaboration
    tier: 2
    docs_url: https://discord.com/developers/docs/topics/gateway-events
    notes: |
      Gateway events are sent over a websocket, not classic HTTP webhooks. Document them anyway
      since the row schema accommodates `delivery: websocket` and the developer-search intent is
      identical ("what events does discord emit"). Also document interaction webhooks separately.
      DECISION (sub 1.5): include Gateway events AND interaction webhooks. Mark `delivery` per
      row.

  - slug: pagerduty
    display_name: PagerDuty
    category: ops
    tier: 2
    docs_url: https://developer.pagerduty.com/docs/db0fa8c8984fc-overview
    notes: |
      Webhook v3 docs enumerate event types with payload examples. Reasonably structured but
      prose-heavy; LLM extraction will outperform a custom parser.

  - slug: clickup
    display_name: ClickUp
    category: collaboration
    tier: 2
    docs_url: https://clickup.com/api/developer-portal/webhooks/
    notes: |
      Webhook docs list events in a single page but payload schemas are sparse in docs (often
      "see object reference"). LLM with extraction-confidence gate.

  - slug: greenhouse
    display_name: Greenhouse
    category: ats
    tier: 2
    docs_url: https://developers.greenhouse.io/webhooks.html
    notes: |
      Webhook docs page lists each event with full payload examples in prose. LLM extraction
      with verification against the Harvest API resource shapes.

  - slug: mailchimp
    display_name: Mailchimp
    category: marketing
    tier: 2
    docs_url: https://mailchimp.com/developer/marketing/guides/sync-audience-data-webhooks/
    notes: |
      Marketing API webhook docs are guide-style prose with embedded payload examples.
      LLM extraction.

  - slug: twilio
    display_name: Twilio
    category: communications
    tier: 2
    docs_url: https://www.twilio.com/docs/usage/webhooks
    notes: |
      Twilio uses webhooks heavily but events are per-product (SMS status callbacks, Voice TwiML
      callbacks, Conversations events, Studio events) with different shapes.
      DECISION (sub 1.5): include Conversations events + Messaging status callbacks. Defer Voice
      TwiML callbacks (request/response, not event-driven) and Studio (low-intent niche) to v2.

  - slug: zoom
    display_name: Zoom
    category: collaboration
    tier: 2
    docs_url: https://developers.zoom.us/docs/api/webhooks/
    notes: |
      Zoom event types are enumerated in a marketplace docs page; per-event payload schemas are
      on individual reference pages. Two-step crawl; LLM-assisted on schema extraction.

  - slug: loom
    display_name: Loom
    category: collaboration
    tier: 2
    docs_url: https://dev.loom.com/docs/embed-sdk-docs/reference/webhook-events
    notes: |
      Smaller surface. LLM extraction.

  - slug: front
    display_name: Front
    category: support
    tier: 2
    docs_url: https://dev.frontapp.com/docs/webhooks-1
    notes: |
      Webhook events listed in dev portal. LLM extraction; payload schemas referenced rather
      than inlined.

  - slug: help-scout
    display_name: Help Scout
    category: support
    tier: 2
    docs_url: https://developer.helpscout.com/webhooks/
    notes: |
      Webhook events documented per resource (conversation events, customer events). LLM
      extraction.

  - slug: freshdesk
    display_name: Freshdesk
    category: support
    tier: 2
    docs_url: https://developers.freshdesk.com/api/#webhooks
    notes: |
      Webhook automation rules surface — events are user-configured triggers rather than a fixed
      catalog. Extract the set of supported trigger conditions as `event_name`, and the ticket/
      contact object schema as payload shape.

  - slug: pipedrive
    display_name: Pipedrive
    category: crm
    tier: 2
    docs_url: https://pipedrive.readme.io/docs/guide-for-webhooks
    notes: |
      Webhooks docs enumerate event triggers (action × object). LLM extraction; payload shape is
      uniform per object type.

  - slug: close
    display_name: Close
    category: crm
    tier: 2
    docs_url: https://developer.close.com/topics/webhooks
    notes: |
      Webhook subscriptions reference. LLM extraction.

  - slug: attio
    display_name: Attio
    category: crm
    tier: 2
    docs_url: https://developers.attio.com/docs/webhooks
    notes: |
      Younger product, smaller event surface. LLM extraction straightforward.

  - slug: bamboohr
    display_name: BambooHR
    category: hris
    tier: 2
    docs_url: https://documentation.bamboohr.com/docs/getting-started-with-webhooks
    notes: |
      Webhook events documented in guide form. LLM extraction.

  - slug: gusto
    display_name: Gusto
    category: hris
    tier: 2
    docs_url: https://docs.gusto.com/embedded-payroll/docs/webhooks-overview
    notes: |
      Swapped in for Rippling per operator decision 2026-05-13 (resolution of sub 1.2 ToS audit
      inconclusive verdict on Rippling). Gusto Embedded webhook docs are publicly accessible
      without login. ToS verdict: provisionally cleared on the same posture as the rest of the
      audit; full ToS read deferred until the Gusto extractor lands in sub 1.5.

  - slug: ashby
    display_name: Ashby
    category: ats
    tier: 2
    docs_url: https://developers.ashbyhq.com/docs/webhooks
    notes: |
      ATS webhook events well-documented. LLM extraction.

# alternates (held in reserve for sub 1.2 fallout if any vendor above gets dropped on ToS grounds)
alternates:
  - slug: rippling
    display_name: Rippling
    category: hris
    docs_url: https://developer.rippling.com/docs/rippling-api/branches/master/docs/Webhooks/Webhooks.md
    notes: dropped from v1 in sub 1.2 (ToS audit inconclusive; developer portal may require login, conflicts with operator-no-accounts policy); held for v2 if a public-docs surface emerges

  - slug: lever
    display_name: Lever
    category: ats
    docs_url: https://hire.lever.co/developer/documentation
    notes: backup for any ATS slot

  - slug: shopify
    display_name: Shopify
    category: ecommerce
    docs_url: https://shopify.dev/docs/api/webhooks
    notes: out-of-category but huge event surface; hold until v2

  - slug: airtable
    display_name: Airtable
    category: collaboration
    docs_url: https://airtable.com/developers/web/api/model/webhooks-notification
    notes: backup collaboration slot