> ## Documentation Index > Fetch the complete documentation index at: https://docs.hyperspell.com/llms.txt > Use this file to discover all available pages before exploring further. # Describe a source's full live query surface > Return the source's capability manifest: capabilities, the raw read surface, and guidance — what the caller can compose a query against. ## OpenAPI ````yaml https://app.stainlessapi.com/api/spec/documented/hyperspell.yaml get /live/{source}/manifest openapi: 3.1.0 info: title: Hyperspell API summary: >- Hyperspell is the memory layer for AI apps and agents. Through the API, you can add memories, connect data sources to index them in real-time, and search memories with our natural language query engine. termsOfService: https://hyperspell.com/blog/tos contact: name: Hyperspell url: https://hyperspell.com/ email: hello@hyperspell.com version: 0.32.1 servers: - url: https://api.hyperspell.com description: Production security: - APIKey: [] AsUser: [] paths: /live/{source}/manifest: get: tags: - Live summary: Describe a source's full live query surface description: >- Return the source's capability manifest: capabilities, the raw read surface, and guidance — what the caller can compose a query against. operationId: live_source_manifest_live__source__manifest_get parameters: - name: source in: path required: true schema: $ref: '#/components/schemas/DocumentProviders' - name: connection_id in: query required: false schema: anyOf: - type: string - type: 'null' description: Specific connection id. title: Connection Id description: Specific connection id. responses: '200': description: Successful Response content: application/json: schema: $ref: '#/components/schemas/SourceManifestResponse' '422': description: Validation Error content: application/json: schema: $ref: '#/components/schemas/HTTPValidationError' security: - APIKey: [] - AsUser: [] components: schemas: DocumentProviders: type: string enum: - reddit - notion - slack - google_calendar - google_mail - box - dropbox - github - google_drive - vault - web_crawler - trace - microsoft_teams - gmail_actions - granola - fathom - fireflies - linear - hubspot - salesforce - coda - lightfield - gong - pylon - clickup title: DocumentProviders SourceManifestResponse: properties: source: $ref: '#/components/schemas/DocumentProviders' auth_provider: anyOf: - $ref: '#/components/schemas/AuthProvider' - type: 'null' capabilities: items: $ref: '#/components/schemas/Capability' type: array title: Capabilities raw_read_ops: items: $ref: '#/components/schemas/RawReadOp' type: array title: Raw Read Ops raw_discovery: type: string title: Raw Discovery guidance: items: type: string type: array title: Guidance resolvable: items: type: string type: array title: Resolvable type: object required: - source - auth_provider - capabilities - raw_read_ops - raw_discovery title: SourceManifestResponse description: >- Everything the caller can ask of a source — read before composing a query. HTTPValidationError: properties: detail: items: $ref: '#/components/schemas/ValidationError' type: array title: Detail type: object title: HTTPValidationError AuthProvider: type: string enum: - nango - unified - whitelabel title: AuthProvider Capability: type: string enum: - list_resources - fetch_resource - search_live - passthrough - resolve - query_structured title: Capability description: >- Live-access capabilities an integration may implement. Used by LiveSourceAccess (and its LLM/HTTP surfaces) to advertise what a connected source can do on-demand. LIST_RESOURCES / FETCH_RESOURCE / SEARCH_LIVE are detected via method-override introspection (see Integration.supports), so they reflect reality with no per-integration declaration to keep in sync. PASSTHROUGH is the exception: it is *declared* (via ``live_raw_read_ops``), not introspected — see specs/components/live-query-framework.md. A source supports it when it exposes a non-empty raw read surface (the agent composes raw provider reads against it). RawReadOp: properties: name: type: string title: Name method: type: string enum: - GET - POST title: Method default: GET path: type: string title: Path description: type: string title: Description params: additionalProperties: type: string type: object title: Params graphql: type: boolean title: Graphql default: false type: object required: - name - path - description title: RawReadOp description: >- One raw read operation a source exposes through live passthrough. A curated (or, later, schema-derived) description of a provider-native READ endpoint the agent may compose against — never a write. ``path`` is the provider-native path or prefix passed to ``Integration.proxy`` (e.g. ``gmail/v1/users/me/messages``); a request matches when its path starts with ``path`` and its method equals ``method``. ``params`` documents useful query params (name -> description) for the agent, and is not validated. ``graphql=True`` marks a read-via-POST GraphQL endpoint: method alone can't keep it read-only (mutations also POST here), so the request body's GraphQL document is checked to be a query — ``mutation``/``subscription`` are rejected. ValidationError: properties: loc: items: anyOf: - type: string - type: integer type: array title: Location msg: type: string title: Message type: type: string title: Error Type type: object required: - loc - msg - type title: ValidationError securitySchemes: APIKey: type: http description: >- API Key or JWT User Token. If using an API Key, set the X-As-User header to act as a specific user. A JWT User Token is always scoped to a specific user. scheme: bearer bearerFormat: Bearer AsUser: type: apiKey description: >- Optionally set this header to act as a specific user when using an API Key, equivalent to first exchanging the API Key for a User Token in: header name: X-As-User ````