generated: '2026-08-13'
method: derived
source: >-
  npm @bigfootai/noded-sdk@0.1.2 (dist/index.d.ts types, dist/index.js selection sets) +
  https://www.getnoded.ai/developers/AGENTS.md
note: >-
  Derived from the published first-party SDK's exported types and the provider's own AGENTS.md
  description of the model ("People and Accounts are both tags"; "Activity is a timeline of
  blocks"; "Memory is what the graph has learned, topic to value"). No OpenAPI exists and the
  GraphQL endpoint is not anonymously reachable, so the underlying schema types (Tag, Block,
  Note, UserMemory) are named only as far as the SDK names them. This is the SDK's flattened
  projection, not the full graph schema.
domain: customer context graph
id_style: opaque string (GraphQL _id, surfaced by the SDK as id)
entities:
  - name: Person
    schema_type: Tag
    description: A contact in the customer graph. A Tag that carries an email.
    fields:
      - {name: id, type: string, required: true}
      - {name: name, type: string, required: true}
      - {name: email, type: string, required: false}
      - {name: emailDomain, type: string, required: false}
      - {name: avatar, type: string, required: false}
      - {name: description, type: string, required: false}
      - {name: externalId, type: string, required: false, note: id in the source system of record}
      - {name: raw, type: unknown, required: false, note: the untouched underlying Tag}
  - name: Account
    schema_type: Tag
    description: A company/organization in the customer graph. A Tag that carries a domain.
    fields:
      - {name: id, type: string, required: true}
      - {name: name, type: string, required: true}
      - {name: domain, type: string, required: false}
      - {name: avatar, type: string, required: false}
      - {name: description, type: string, required: false}
      - {name: externalId, type: string, required: false}
      - {name: raw, type: unknown, required: false}
  - name: ActivityItem
    schema_type: Block
    description: >-
      A unit of work on the timeline — note, email, task, message, meeting transcription, or
      event. Noded's pricing meters these as "blocks".
    fields:
      - {name: id, type: string, required: true}
      - {name: type, type: string, required: true}
      - {name: title, type: string, required: false}
      - {name: preview, type: string, required: false}
      - {name: dateCreated, type: number, required: false, note: epoch}
      - {name: dateUpdated, type: number, required: false, note: epoch}
      - {name: raw, type: unknown, required: false}
  - name: Note
    schema_type: Note
    description: A note written back into the graph. A Block subtype.
    fields:
      - {name: id, type: string, required: true}
      - {name: title, type: string, required: false}
      - {name: document, type: string, required: false, note: Noded editor document, JSON string}
      - {name: favorite, type: boolean, required: false}
      - {name: raw, type: unknown, required: false}
  - name: MemoryEntry
    schema_type: UserMemory
    description: >-
      What the graph has learned about a person or account. The SDK flattens it to
      { id, content }; the underlying UserMemory carries entries[] of
      { topic, value, confidence, source, dateCreated, dateUpdated }.
    fields:
      - {name: id, type: string, required: true}
      - {name: content, type: string, required: true}
      - {name: 'entries[].topic', type: string, required: false}
      - {name: 'entries[].value', type: string, required: false}
      - {name: 'entries[].confidence', type: string, required: false}
      - {name: 'entries[].source', type: string, required: false}
      - {name: 'entries[].dateCreated', type: number, required: false}
      - {name: 'entries[].dateUpdated', type: number, required: false}
relationships:
  - from: Person
    to: ActivityItem
    kind: has_many
    via: tagId
    operation: 'blocks(input: FindBlocksInput)'
  - from: Account
    to: ActivityItem
    kind: has_many
    via: tagId
    operation: 'blocks(input: FindBlocksInput)'
  - from: Person
    to: MemoryEntry
    kind: has_many
    via: tagId
    operation: 'userMemoriesForTag(input: FindUserMemoriesForTagInput)'
  - from: Account
    to: MemoryEntry
    kind: has_many
    via: tagId
    operation: 'userMemoriesForTag(input: FindUserMemoriesForTagInput)'
  - from: Note
    to: ActivityItem
    kind: is_a
    via: Block
    note: notes.list() is blocks() filtered to notes — a Note is a Block
  - from: Person
    to: Account
    kind: inferred
    via: emailDomain -> domain
    note: >-
      Not a declared edge in the SDK. The provider's own quickstart resolves an account by email
      domain (people.search('acme.com')), so domain matching is the documented join in practice.
  - from: Person
    to: external system record
    kind: has_one
    via: externalId
  - from: Account
    to: external system record
    kind: has_one
    via: externalId
entity_count: 5
relationship_count: 8
underlying_schema_types:
  - Tag
  - Block
  - Note
  - UserMemory
gaps:
  - >-
    Person and Account are the same underlying Tag type; the SDK does not expose the
    discriminator that separates them, so a consumer cannot tell from the payload alone.
  - No object reference or id-prefix documentation is published.