QuantCDN AI Vector Database API

openapi: 3.2.0
info:
  description: Unified API for QuantCDN Admin and QuantCloud Platform services
  title: QuantCDN AI Agents AI Vector Database API
  version: 4.15.8
servers:
- description: QuantCDN Public Cloud
  url: https://dashboard.quantcdn.io
- description: QuantGov Cloud
  url: https://dash.quantgov.cloud
security:
- BearerAuth: []
tags:
- description: Vector database collections for RAG and semantic search
  name: AI Vector Database
paths:
  /api/v3/organizations/{organisation}/ai/vector-db/collections:
    get:
      description: Lists all vector database collections (knowledge bases) for an organization.
      operationId: listVectorCollections
      parameters:
      - description: The organisation ID
        explode: false
        in: path
        name: organisation
        required: true
        schema:
          type: string
        style: simple
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/listVectorCollections_200_response'
          description: Collections retrieved successfully
        '403':
          description: Access denied
        '500':
          description: Failed to retrieve collections
      summary: List Vector Database Collections
      tags:
      - AI Vector Database
    post:
      description: "Creates a new vector database collection (knowledge base category) for semantic search. Collections store documents with embeddings for RAG (Retrieval Augmented Generation).\n     *\n     * **Use Cases:**\n     * - Product documentation ('docs')\n     * - Company policies ('policies')\n     * - Support knowledge base ('support')\n     * - Technical specifications ('specs')"
      operationId: createVectorCollection
      parameters:
      - description: The organisation ID
        explode: false
        in: path
        name: organisation
        required: true
        schema:
          type: string
        style: simple
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/createVectorCollection_request'
        required: true
      responses:
        '201':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/createVectorCollection_201_response'
          description: Collection created successfully
        '400':
          description: Invalid request parameters
        '403':
          description: Access denied
        '409':
          description: Collection with this name already exists
        '500':
          description: Failed to create collection
      summary: Create Vector Database Collection
      tags:
      - AI Vector Database
  /api/v3/organizations/{organisation}/ai/vector-db/collections/{collectionId}:
    delete:
      description: Deletes a vector database collection and all its documents. This action cannot be undone.
      operationId: deleteVectorCollection
      parameters:
      - description: The organisation ID
        explode: false
        in: path
        name: organisation
        required: true
        schema:
          type: string
        style: simple
      - description: The collection ID
        explode: false
        in: path
        name: collectionId
        required: true
        schema:
          format: uuid
          type: string
        style: simple
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/deleteSkillCollection_200_response'
          description: Collection deleted successfully
        '403':
          description: Access denied
        '404':
          description: Collection not found
        '500':
          description: Failed to delete collection
      summary: Delete Collection
      tags:
      - AI Vector Database
    get:
      description: Get detailed information about a specific vector database collection.
      operationId: getVectorCollection
      parameters:
      - description: The organisation ID
        explode: false
        in: path
        name: organisation
        required: true
        schema:
          type: string
        style: simple
      - description: The collection ID
        explode: false
        in: path
        name: collectionId
        required: true
        schema:
          format: uuid
          type: string
        style: simple
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/getVectorCollection_200_response'
          description: Collection details retrieved successfully
        '403':
          description: Access denied
        '404':
          description: Collection not found
        '500':
          description: Failed to retrieve collection
      summary: Get Collection Details
      tags:
      - AI Vector Database
  /api/v3/organizations/{organisation}/ai/vector-db/collections/{collectionId}/documents:
    delete:
      description: "Delete documents from a collection. Supports three deletion modes:\n     *\n     * 1. **Purge All** - Set `purgeAll: true` to delete ALL documents in the collection\n     *\n     * 2. **By Document IDs** - Provide `documentIds` array with specific document UUIDs\n     *\n     * 3. **By Metadata** - Provide `metadata` object with `field` and `values` to delete documents where the metadata field matches any of the values\n     *\n     * **Drupal Integration:**\n     * When using with Drupal AI Search, use metadata deletion with:\n     * - `field: 'drupal_entity_id'` to delete all chunks for specific entities\n     * - `field: 'drupal_long_id'` to delete specific chunks"
      operationId: deleteVectorDocuments
      parameters:
      - description: Organisation machine name
        explode: false
        in: path
        name: organisation
        required: true
        schema:
          type: string
        style: simple
      - description: Collection UUID
        explode: false
        in: path
        name: collectionId
        required: true
        schema:
          format: uuid
          type: string
        style: simple
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/deleteVectorDocuments_request'
        required: true
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/deleteVectorDocuments_200_response'
          description: Documents deleted successfully
        '400':
          description: Invalid request - must specify purgeAll, documentIds, or metadata
        '403':
          description: Access denied
        '404':
          description: Collection not found
        '500':
          description: Failed to delete documents
      summary: Delete Documents from Collection
      tags:
      - AI Vector Database
    get:
      description: Lists documents in a collection with pagination. Supports filtering by document key.
      operationId: listVectorDocuments
      parameters:
      - explode: false
        in: path
        name: organisation
        required: true
        schema:
          type: string
        style: simple
      - explode: false
        in: path
        name: collectionId
        required: true
        schema:
          format: uuid
          type: string
        style: simple
      - description: Filter by document key
        explode: true
        in: query
        name: key
        required: false
        schema:
          type: string
        style: form
      - explode: true
        in: query
        name: limit
        required: false
        schema:
          default: 50
          maximum: 100
          type: integer
        style: form
      - explode: true
        in: query
        name: offset
        required: false
        schema:
          default: 0
          type: integer
        style: form
      responses:
        '200':
          description: Documents retrieved successfully
        '403':
          description: Access denied
        '404':
          description: Collection not found
        '500':
          description: Failed to list documents
      summary: List Documents in Collection
      tags:
      - AI Vector Database
    post:
      description: "Uploads documents to a vector database collection with automatic embedding generation. Documents are chunked (if needed), embedded using the collection's embedding model, and stored.\n     *\n     * **Supported Content:**\n     * - Plain text content\n     * - URLs to fetch content from\n     * - Markdown documents\n     *\n     * **Metadata:**\n     * Each document can include metadata (title, source_url, section, tags) that is returned with search results."
      operationId: uploadVectorDocuments
      parameters:
      - description: The organisation ID
        explode: false
        in: path
        name: organisation
        required: true
        schema:
          type: string
        style: simple
      - description: The collection ID
        explode: false
        in: path
        name: collectionId
        required: true
        schema:
          format: uuid
          type: string
        style: simple
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/uploadVectorDocuments_request'
        required: true
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/uploadVectorDocuments_200_response'
          description: Documents uploaded successfully
        '400':
          description: Invalid request parameters
        '403':
          description: Access denied
        '404':
          description: Collection not found
        '500':
          description: Failed to upload documents
      summary: Upload Documents to Collection
      tags:
      - AI Vector Database
  /api/v3/organizations/{organisation}/ai/vector-db/collections/{collectionId}/query:
    post:
      description: "Performs semantic search on a collection using vector similarity. Returns the most relevant documents based on meaning, not keyword matching.\n     *\n     * **Three Search Modes:**\n     *\n     * 1. **Text Query** - Provide `query` string, server generates embedding\n     *    - Query text is embedded using the collection's embedding model\n     *    - Embeddings are cached for repeated queries\n     *\n     * 2. **Vector Query** - Provide pre-computed `vector` array\n     *    - Skip embedding generation (faster)\n     *    - Useful when you've already embedded the query elsewhere\n     *    - Vector dimension must match collection (e.g., 1024 for Titan v2)\n     *\n     * 3. **Metadata List** - Set `listByMetadata: true` with `filter`\n     *    - Skip semantic search entirely\n     *    - Return all documents matching the filter\n     *    - Supports cursor-based pagination for large datasets\n     *    - Results ordered by sortBy/sortOrder (default: created_at DESC)\n     *\n     * **Filtering:**\n     * - `filter.exact`: Exact match on metadata fields (AND logic)\n     * - `filter.contains`: Array contains filter for tags (ANY match)\n     * - Filters can be combined with semantic search or used alone with listByMetadata\n     *\n     * **Pagination (listByMetadata mode only):**\n     * - Use `cursor` from previous response's `nextCursor` to get next page\n     * - Uses keyset pagination for efficient traversal of large datasets\n     * - Control sort with `sortBy` and `sortOrder`\n     *\n     * **Use Cases:**\n     * - Find relevant documentation for user questions\n     * - Power RAG (Retrieval Augmented Generation) in AI assistants\n     * - Semantic search across knowledge bases\n     * - List all artifacts by building/worker/tag"
      operationId: queryVectorCollection
      parameters:
      - description: The organisation ID
        explode: false
        in: path
        name: organisation
        required: true
        schema:
          type: string
        style: simple
      - description: The collection ID
        explode: false
        in: path
        name: collectionId
        required: true
        schema:
          format: uuid
          type: string
        style: simple
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/queryVectorCollection_request'
        required: true
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/queryVectorCollection_200_response'
          description: Search completed successfully
        '400':
          description: Invalid request parameters
        '403':
          description: Access denied
        '404':
          description: Collection not found
        '500':
          description: Failed to perform search
      summary: Semantic Search Query
      tags:
      - AI Vector Database
components:
  schemas:
    deleteVectorDocuments_request_metadata:
      properties:
        field:
          description: Metadata field name (e.g., 'drupal_entity_id')
          type: string
        values:
          description: Values to match (OR logic)
          items:
            type: string
          type: array
      type: object
    getVectorCollection_200_response:
      example:
        collection:
          documentCount: 0
          createdAt: 2000-01-23 04:56:07+00:00
          name: name
          description: description
          embeddingModel: embeddingModel
          collectionId: 046b6c7f-0b8a-43b9-b35d-6489e6daee91
          dimensions: 6
          updatedAt: 2000-01-23 04:56:07+00:00
      properties:
        collection:
          $ref: '#/components/schemas/getVectorCollection_200_response_collection'
      type: object
    createVectorCollection_201_response_collection:
      example:
        name: name
        description: description
        embeddingModel: embeddingModel
        collectionId: 046b6c7f-0b8a-43b9-b35d-6489e6daee91
        dimensions: 0
      properties:
        collectionId:
          format: uuid
          type: string
        name:
          type: string
        description:
          type: string
        embeddingModel:
          type: string
        dimensions:
          type: integer
      type: object
    deleteVectorDocuments_200_response:
      example:
        deletedCount: 0
        message: message
        collectionId: collectionId
      properties:
        message:
          type: string
        collectionId:
          type: string
        deletedCount:
          type: integer
      type: object
    uploadVectorDocuments_request:
      properties:
        documents:
          items:
            $ref: '#/components/schemas/uploadVectorDocuments_request_documents_inner'
          type: array
      required:
      - documents
      type: object
    getVectorCollection_200_response_collection:
      example:
        documentCount: 0
        createdAt: 2000-01-23 04:56:07+00:00
        name: name
        description: description
        embeddingModel: embeddingModel
        collectionId: 046b6c7f-0b8a-43b9-b35d-6489e6daee91
        dimensions: 6
        updatedAt: 2000-01-23 04:56:07+00:00
      properties:
        collectionId:
          format: uuid
          type: string
        name:
          type: string
        description:
          type: string
        documentCount:
          type: integer
        embeddingModel:
          type: string
        dimensions:
          type: integer
        createdAt:
          format: date-time
          type: string
        updatedAt:
          format: date-time
          type: string
      type: object
    uploadVectorDocuments_200_response:
      example:
        chunksCreated: 0
        success: true
        message: message
        documentIds:
        - documentIds
        - documentIds
      properties:
        success:
          example: true
          type: boolean
        documentIds:
          items:
            type: string
          type: array
        chunksCreated:
          type: integer
        message:
          type: string
      type: object
    queryVectorCollection_request:
      properties:
        query:
          description: Natural language search query (mutually exclusive with vector)
          example: How do I authenticate with the API?
          maxLength: 1000
          minLength: 3
          type: string
        vector:
          description: Pre-computed embedding vector (mutually exclusive with query). Array length must match collection dimension.
          example:
          - 0.0234
          - -0.0891
          - 0.0456
          items:
            format: float
            type: number
          type: array
        limit:
          default: 5
          description: Maximum number of results to return
          maximum: 20
          minimum: 1
          type: integer
        threshold:
          default: 0.7
          description: Minimum similarity score (0-1, higher = more relevant)
          format: float
          maximum: 1
          minimum: 0
          type: number
        includeEmbeddings:
          default: false
          description: Include embedding vectors in response (for debugging)
          type: boolean
        filter:
          $ref: '#/components/schemas/queryVectorCollection_request_filter'
        listByMetadata:
          default: false
          description: If true, skip semantic search and return all documents matching the filter. Requires filter. Supports cursor pagination.
          type: boolean
        cursor:
          description: Pagination cursor for listByMetadata mode. Use nextCursor from previous response. Opaque format - do not construct manually.
          type: string
        sortBy:
          default: created_at
          description: Field to sort by in listByMetadata mode
          enum:
          - created_at
          - document_id
          type: string
        sortOrder:
          default: desc
          description: Sort direction in listByMetadata mode
          enum:
          - asc
          - desc
          type: string
      type: object
    createVectorCollection_request:
      properties:
        name:
          description: Collection name (used for reference)
          example: product-documentation
          type: string
        description:
          example: Product user guides and API documentation
          type: string
        embeddingModel:
          description: 'Embedding model to use. Supported: amazon.titan-embed-text-v2:0, cohere.embed-english-v3, cohere.embed-multilingual-v3'
          example: amazon.titan-embed-text-v2:0
          type: string
        dimensions:
          description: 'Embedding dimensions (default: 1024)'
          example: 1024
          type: integer
      required:
      - embeddingModel
      - name
      type: object
    listVectorCollections_200_response:
      example:
        collections:
        - documentCount: 0
          createdAt: 2000-01-23 04:56:07+00:00
          name: product-docs
          description: description
          embeddingModel: amazon.titan-embed-text-v2:0
          collectionId: 046b6c7f-0b8a-43b9-b35d-6489e6daee91
        - documentCount: 0
          createdAt: 2000-01-23 04:56:07+00:00
          name: product-docs
          description: description
          embeddingModel: amazon.titan-embed-text-v2:0
          collectionId: 046b6c7f-0b8a-43b9-b35d-6489e6daee91
        count: 6
      properties:
        collections:
          items:
            $ref: '#/components/schemas/listVectorCollections_200_response_collections_inner'
          type: array
        count:
          type: integer
      type: object
    deleteVectorDocuments_request:
      properties:
        purgeAll:
          description: Delete ALL documents in collection
          type: boolean
        documentIds:
          description: Delete specific documents by UUID
          items:
            format: uuid
            type: string
          type: array
        keys:
          description: Delete documents by key
          items:
            maxLength: 512
            type: string
          type: array
        metadata:
          $ref: '#/components/schemas/deleteVectorDocuments_request_metadata'
      type: object
    uploadVectorDocuments_request_documents_inner:
      properties:
        content:
          description: Document text content
          type: string
        key:
          description: Stable document key for upsert
          maxLength: 512
          type: string
        metadata:
          $ref: '#/components/schemas/uploadVectorDocuments_request_documents_inner_metadata'
      required:
      - content
      type: object
    createVectorCollection_201_response:
      example:
        success: true
        collection:
          name: name
          description: description
          embeddingModel: embeddingModel
          collectionId: 046b6c7f-0b8a-43b9-b35d-6489e6daee91
          dimensions: 0
        message: Collection created successfully
      properties:
        success:
          example: true
          type: boolean
        collection:
          $ref: '#/components/schemas/createVectorCollection_201_response_collection'
        message:
          example: Collection created successfully
          type: string
      type: object
    queryVectorCollection_200_response_pagination:
      description: Pagination info (listByMetadata mode only)
      example:
        sortOrder: asc
        limit: 5
        sortBy: created_at
      properties:
        sortBy:
          enum:
          - created_at
          - document_id
          type: string
        sortOrder:
          enum:
          - asc
          - desc
          type: string
        limit:
          type: integer
      type:
      - object
      - 'null'
    queryVectorCollection_request_filter:
      description: Filter results by metadata fields. Applied AFTER semantic search (or alone in listByMetadata mode). All conditions use AND logic.
      properties:
        exact:
          additionalProperties: true
          description: Exact match on metadata fields. Keys are metadata field names, values are expected values.
          example:
            buildingId: building-123
            type: research-report
          type: object
        contains:
          additionalProperties:
            items:
              type: string
            type: array
          description: Array contains filter for array metadata fields (like tags). Returns documents where the metadata array contains ANY of the specified values.
          example:
            tags:
            - important
            - reviewed
          type: object
      type: object
    deleteSkillCollection_200_response:
      example:
        success: true
        message: Collection deleted successfully
      properties:
        success:
          example: true
          type: boolean
        message:
          example: Collection deleted successfully
          type: string
      type: object
    queryVectorCollection_200_response_results_inner:
      example:
        metadata:
          key: ''
        similarity: 0.08008282
        documentId: 046b6c7f-0b8a-43b9-b35d-6489e6daee91
        embedding:
        - 6.027456183070403
        - 6.027456183070403
        content: content
      properties:
        documentId:
          format: uuid
          type: string
        content:
          description: Document text content
          type: string
        similarity:
          description: Cosine similarity score (1.0 for metadata-only queries)
          format: float
          maximum: 1
          minimum: 0
          type: number
        metadata:
          additionalProperties: true
          type: object
        embedding:
          description: Vector embedding (only if includeEmbeddings=true)
          items:
            type: number
          type: array
      type: object
    queryVectorCollection_200_response:
      example:
        filter: '{}'
        nextCursor: nextCursor
        pagination:
          sortOrder: asc
          limit: 5
          sortBy: created_at
        query: query
        executionTimeMs: 5
        searchMode: text
        count: 1
        hasMore: true
        results:
        - metadata:
            key: ''
          similarity: 0.08008282
          documentId: 046b6c7f-0b8a-43b9-b35d-6489e6daee91
          embedding:
          - 6.027456183070403
          - 6.027456183070403
          content: content
        - metadata:
            key: ''
          similarity: 0.08008282
          documentId: 046b6c7f-0b8a-43b9-b35d-6489e6daee91
          embedding:
          - 6.027456183070403
          - 6.027456183070403
          content: content
        collectionId: 046b6c7f-0b8a-43b9-b35d-6489e6daee91
      properties:
        results:
          items:
            $ref: '#/components/schemas/queryVectorCollection_200_response_results_inner'
          type: array
        query:
          description: Original query text (null if vector or metadata search was used)
          type:
          - string
          - 'null'
        searchMode:
          description: 'Search mode used: text (query provided), vector (pre-computed), metadata (listByMetadata)'
          enum:
          - text
          - vector
          - metadata
          type: string
        filter:
          description: Filter that was applied (if any)
          type:
          - object
          - 'null'
        count:
          description: Number of results returned
          type: integer
        executionTimeMs:
          description: Query execution time in milliseconds
          type: integer
        collectionId:
          format: uuid
          type: string
        hasMore:
          description: True if more results available (listByMetadata mode only)
          type: boolean
        nextCursor:
          description: Cursor for next page. Pass as cursor param to continue. Null when no more results. Only in listByMetadata mode.
          type:
          - string
          - 'null'
        pagination:
          $ref: '#/components/schemas/queryVectorCollection_200_response_pagination'
      type: object
    uploadVectorDocuments_request_documents_inner_metadata:
      properties:
        title:
          type: string
        source_url:
          type: string
        section:
          type: string
        tags:
          items:
            type: string
          type: array
      type: object
    listVectorCollections_200_response_collections_inner:
      example:
        documentCount: 0
        createdAt: 2000-01-23 04:56:07+00:00
        name: product-docs
        description: description
        embeddingModel: amazon.titan-embed-text-v2:0
        collectionId: 046b6c7f-0b8a-43b9-b35d-6489e6daee91
      properties:
        collectionId:
          format: uuid
          type: string
        name:
          example: product-docs
          type: string
        description:
          type: string
        documentCount:
          type: integer
        embeddingModel:
          example: amazon.titan-embed-text-v2:0
          type: string
        createdAt:
          format: date-time
          type: string
      type: object
  securitySchemes:
    BearerAuth:
      bearerFormat: JWT
      description: 'Enter your Bearer token in the format: `Bearer <your-token-here>`. Obtain your API token from the QuantCDN dashboard under Profile > API Tokens.'
      scheme: bearer
      type: http