Sarvam documentIntelligence API

openapi: 3.1.0
info:
  title: Endpoints chat documentIntelligence API
  version: 1.0.0
servers:
- url: https://api.sarvam.ai
  description: Production
tags:
- name: documentIntelligence
paths:
  /doc-digitization/job/v1:
    post:
      operationId: initialise
      summary: Create Document Intelligence Job
      description: 'Creates a new Document Intelligence job.


        **Supported Languages (BCP-47 format):**

        - `hi-IN`: Hindi (default)

        - `en-IN`: English

        - `bn-IN`: Bengali

        - `gu-IN`: Gujarati

        - `kn-IN`: Kannada

        - `ml-IN`: Malayalam

        - `mr-IN`: Marathi

        - `od-IN`: Odia

        - `pa-IN`: Punjabi

        - `ta-IN`: Tamil

        - `te-IN`: Telugu

        - `ur-IN`: Urdu

        - `as-IN`: Assamese

        - `brx-IN`: Bodo

        - `doi-IN`: Dogri

        - `ks-IN`: Kashmiri

        - `kok-IN`: Konkani

        - `mai-IN`: Maithili

        - `mni-IN`: Manipuri

        - `ne-IN`: Nepali

        - `sa-IN`: Sanskrit

        - `sat-IN`: Santali

        - `sd-IN`: Sindhi


        **Output Formats (delivered as ZIP file):**

        - `html`: Structured HTML files with layout preservation

        - `md`: Markdown files (default)

        - `json`: Structured JSON files for programmatic processing'
      tags:
      - documentIntelligence
      parameters:
      - name: api-subscription-key
        in: header
        required: true
        schema:
          type: string
      responses:
        '202':
          description: Successful Response
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationCreateJobResponse'
        '400':
          description: Bad Request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '403':
          description: Forbidden
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '429':
          description: Quota Exceeded / Rate Limited
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '500':
          description: Internal Server Error
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '503':
          description: Service Unavailable
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/DocDigitizationCreateJobRequest'
  /doc-digitization/job/v1/upload-files:
    post:
      operationId: get-upload-links
      summary: Get Document Intelligence Upload URLs
      description: 'Returns presigned URLs for uploading input files.


        **File Constraints:**

        - Exactly one file required (PDF or ZIP)

        - PDF files: `.pdf` extension

        - ZIP files: `.zip` extension'
      tags:
      - documentIntelligence
      parameters:
      - name: api-subscription-key
        in: header
        required: true
        schema:
          type: string
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationUploadFilesResponse'
        '400':
          description: Bad Request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '403':
          description: Forbidden
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '429':
          description: Quota Exceeded / Rate Limited
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '500':
          description: Internal Server Error
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '503':
          description: Service Unavailable
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/DocDigitizationUploadFilesRequest'
  /doc-digitization/job/v1/{job_id}/start:
    post:
      operationId: start
      summary: Start Document Intelligence Job
      description: 'Validates the uploaded file and starts processing.


        **Validation Checks:**

        - File must be uploaded before starting

        - File size must not exceed 200 MB

        - PDF must be parseable by the PDF parser

        - ZIP must contain only JPEG/PNG images

        - ZIP must be flat (no nested folders beyond one level)

        - ZIP must contain at least one valid image

        - Page/image count must not exceed 10 (returns `422` with `max_page_limit_exceeded` if exceeded)

        - User must have sufficient credits


        **Processing:**

        Job runs asynchronously. Poll the status endpoint or use webhook callback for completion notification.'
      tags:
      - documentIntelligence
      parameters:
      - name: job_id
        in: path
        description: The unique identifier of the job
        required: true
        schema:
          type: string
          format: uuid
      - name: api-subscription-key
        in: header
        required: true
        schema:
          type: string
      responses:
        '202':
          description: Successful Response
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationJobStatusResponse'
        '400':
          description: Bad Request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '403':
          description: Forbidden
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '429':
          description: Quota Exceeded / Rate Limited
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '500':
          description: Internal Server Error
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '503':
          description: Service Unavailable
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
  /doc-digitization/job/v1/{job_id}/status:
    get:
      operationId: get-status
      summary: Get Document Intelligence Job Status
      description: 'Returns the current status of a job with page-level metrics.


        **Job States:**

        - `Accepted`: Job created, awaiting file upload

        - `Pending`: File uploaded, waiting to start

        - `Running`: Processing in progress

        - `Completed`: All pages processed successfully

        - `PartiallyCompleted`: Some pages succeeded, some failed

        - `Failed`: All pages failed or job-level error


        **Page Metrics:**

        Response includes detailed progress: total pages, pages processed, succeeded, failed, and per-page errors.'
      tags:
      - documentIntelligence
      parameters:
      - name: job_id
        in: path
        description: The unique identifier of the job
        required: true
        schema:
          type: string
          format: uuid
      - name: api-subscription-key
        in: header
        required: true
        schema:
          type: string
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationJobStatusResponse'
        '400':
          description: Bad Request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '403':
          description: Forbidden
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '429':
          description: Quota Exceeded / Rate Limited
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '500':
          description: Internal Server Error
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '503':
          description: Service Unavailable
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
  /doc-digitization/job/v1/{job_id}/download-files:
    post:
      operationId: get-download-links
      summary: Get Document Intelligence Download URLs
      description: 'Returns presigned URLs for downloading output files.


        **Method:** POST only. A GET request to this path returns 405 Method Not Allowed.


        **Prerequisites:**

        - Job must be in `Completed` or `PartiallyCompleted` state

        - Failed jobs have no output available'
      tags:
      - documentIntelligence
      parameters:
      - name: job_id
        in: path
        description: The unique identifier of the job
        required: true
        schema:
          type: string
          format: uuid
      - name: api-subscription-key
        in: header
        required: true
        schema:
          type: string
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationDownloadFilesResponse'
        '400':
          description: Bad Request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '403':
          description: Forbidden
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '429':
          description: Quota Exceeded / Rate Limited
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '500':
          description: Internal Server Error
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
        '503':
          description: Service Unavailable
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DocDigitizationErrorMessage'
components:
  schemas:
    DocDigitizationJobDetail:
      type: object
      properties:
        inputs:
          type: array
          items:
            $ref: '#/components/schemas/TaskFileDetails'
          description: Input file(s) for this task
        outputs:
          type: array
          items:
            $ref: '#/components/schemas/TaskFileDetails'
          description: Output file(s) produced
        state:
          $ref: '#/components/schemas/DocDigitizationJobDetailState'
          description: Processing state for this file
        total_pages:
          type: integer
          default: 0
          description: Total pages/images in the input file
        pages_processed:
          type: integer
          default: 0
          description: Number of pages processed so far
        pages_succeeded:
          type: integer
          default: 0
          description: Number of pages successfully processed
        pages_failed:
          type: integer
          default: 0
          description: Number of pages that failed processing
        error_message:
          type: string
          default: ''
          description: Error message if processing failed
        error_code:
          type:
          - string
          - 'null'
          description: Standardized error code if failed
        page_errors:
          type: array
          items:
            $ref: '#/components/schemas/DocDigitizationPageError'
          description: Detailed errors for each failed page
      required:
      - inputs
      - outputs
      - state
      description: Processing details for a single input file with page-level metrics.
      title: DocDigitizationJobDetail
    DocDigitizationErrorDetails:
      type: object
      properties:
        message:
          type: string
          description: Message describing the error
        code:
          $ref: '#/components/schemas/DocDigitizationErrorCode'
          description: Error code for the specific error that has occurred.
        request_id:
          type: string
          default: ''
          description: 'Unique identifier for the request. Format: date_UUID4'
      required:
      - message
      - code
      title: DocDigitizationErrorDetails
    DocDigitizationPageError:
      type: object
      properties:
        page_number:
          type: integer
          description: Page number that failed
        error_code:
          type: string
          description: Standardized error code
        error_message:
          type: string
          description: Human-readable error description
      required:
      - page_number
      - error_code
      - error_message
      description: Error details for a specific page.
      title: DocDigitizationPageError
    DocDigitizationJobState:
      type: string
      enum:
      - Accepted
      - Pending
      - Running
      - Completed
      - PartiallyCompleted
      - Failed
      description: Current state of the document intelligence job
      title: DocDigitizationJobState
    DocDigitizationWebhookCallback:
      type: object
      properties:
        url:
          type: string
          format: uri
          description: HTTPS webhook URL to call upon job completion (HTTP not allowed)
        auth_token:
          type: string
          default: ''
          description: Authorization token sent as X-SARVAM-JOB-CALLBACK-TOKEN header
      required:
      - url
      description: Webhook configuration for job completion notification
      title: DocDigitizationWebhookCallback
    DocDigitizationCreateJobResponse:
      type: object
      properties:
        job_id:
          type: string
          format: uuid
          description: Unique job identifier (UUID)
        storage_container_type:
          $ref: '#/components/schemas/StorageContainerType'
          description: Storage Container Type
        job_parameters:
          $ref: '#/components/schemas/DocDigitizationJobParameters'
          description: '  Job configuration parameters'
        job_state:
          $ref: '#/components/schemas/DocDigitizationJobState'
      required:
      - job_id
      - storage_container_type
      - job_parameters
      - job_state
      title: DocDigitizationCreateJobResponse
    DocDigitizationErrorMessage:
      type: object
      properties:
        error:
          $ref: '#/components/schemas/DocDigitizationErrorDetails'
          description: Error details
      required:
      - error
      title: DocDigitizationErrorMessage
    DocDigitizationJobParameters:
      type: object
      properties:
        language:
          $ref: '#/components/schemas/DocDigitizationSupportedLanguage'
          description: Primary language of the document in BCP-47 format (e.g. en-IN, hi-IN). Use this field name — not language_code (which other Sarvam APIs use). Sending language_code is ignored and defaults to hi-IN.
        output_format:
          $ref: '#/components/schemas/DocDigitizationOutputFormat'
          description: 'Output format for the extracted content (delivered as a ZIP file). Accepted values: md, html, json. Use md for Markdown — not markdown (returns 400).'
      description: Configuration parameters for Document Intelligence job. Specify the document language and desired output format.
      title: DocDigitizationJobParameters
    StorageContainerType:
      type: string
      enum:
      - Azure
      - Local
      - Google
      - Azure_V1
      title: StorageContainerType
    DocDigitizationOutputFormat:
      type: string
      enum:
      - html
      - md
      - json
      description: ' Output format for extracted document content, delivered as a ZIP file. ''html'' returns structured HTML files, ''md'' returns human-readable Markdown files, ''json'' returns structured JSON files for programmatic processing.'
      title: DocDigitizationOutputFormat
    DocDigitizationErrorCode:
      type: string
      enum:
      - invalid_request_error
      - internal_server_error
      - insufficient_quota_error
      - invalid_api_key_error
      - rate_limit_exceeded_error
      - high_load_error
      title: DocDigitizationErrorCode
    DocDigitizationJobStatusResponse:
      type: object
      properties:
        job_id:
          type: string
          format: uuid
          description: Job identifier (UUID)
        job_state:
          $ref: '#/components/schemas/DocDigitizationJobState'
          description: Current job state
        created_at:
          type: string
          format: date-time
          description: Job creation timestamp (ISO 8601)
        updated_at:
          type: string
          format: date-time
          description: Last update timestamp (ISO 8601)
        storage_container_type:
          $ref: '#/components/schemas/StorageContainerType'
          description: Storage backend type
        total_files:
          type: integer
          default: 0
          description: Total input files (always 1)
        successful_files_count:
          type: integer
          default: 0
          description: Files that completed successfully
        failed_files_count:
          type: integer
          default: 0
          description: Files that failed
        error_message:
          type: string
          default: ''
          description: Job-level error message
        job_details:
          type: array
          items:
            $ref: '#/components/schemas/DocDigitizationJobDetail'
          description: Per-file processing details with page metrics
      required:
      - job_id
      - job_state
      - created_at
      - updated_at
      - storage_container_type
      description: Response model for job status endpoint.
      title: DocDigitizationJobStatusResponse
    TaskFileDetails:
      type: object
      properties:
        file_name:
          type: string
        file_id:
          type: string
      required:
      - file_name
      - file_id
      title: TaskFileDetails
    DocDigitizationSupportedLanguage:
      type: string
      enum:
      - hi-IN
      - en-IN
      - bn-IN
      - gu-IN
      - kn-IN
      - ml-IN
      - mr-IN
      - od-IN
      - pa-IN
      - ta-IN
      - te-IN
      - ur-IN
      - as-IN
      - brx-IN
      - doi-IN
      - ks-IN
      - kok-IN
      - mai-IN
      - mni-IN
      - ne-IN
      - sa-IN
      - sat-IN
      - sd-IN
      description: 'BCP-47 language code specifying the primary language of the document. Supports 23 languages: 22 Indian languages (Hindi, Bengali, Tamil, Telugu, Marathi, Gujarati, Kannada, Malayalam, Odia, Punjabi, Assamese, Urdu, Sanskrit, Nepali, Konkani, Maithili, Sindhi, Kashmiri, Dogri, Manipuri, Bodo, Santali) and English language.'
      title: DocDigitizationSupportedLanguage
    DocDigitizationJobDetailState:
      type: string
      enum:
      - Pending
      - Running
      - Success
      - PartialSuccess
      - Failed
      description: Processing state for individual file
      title: DocDigitizationJobDetailState
    DocDigitizationCreateJobRequest:
      type: object
      properties:
        job_parameters:
          $ref: '#/components/schemas/DocDigitizationJobParameters'
          description: Configuration parameters for the Document Intelligence job including language and output format. Defaults to Hindi (hi-IN) and Markdown output if omitted.
        callback:
          oneOf:
          - $ref: '#/components/schemas/DocDigitizationWebhookCallback'
          - type: 'null'
          description: Optional webhook for completion notification
      description: Request body for creating a new document intelligence job
      title: DocDigitizationCreateJobRequest
    FileSignedURLDetails:
      type: object
      properties:
        file_url:
          type: string
        file_metadata:
          type:
          - object
          - 'null'
          additionalProperties:
            description: Any type
      required:
      - file_url
      title: FileSignedURLDetails
    DocDigitizationDownloadFilesResponse:
      type: object
      properties:
        job_id:
          type: string
          format: uuid
          description: Job identifier (UUID)
        job_state:
          $ref: '#/components/schemas/DocDigitizationJobState'
          description: Current job state
        storage_container_type:
          $ref: '#/components/schemas/StorageContainerType'
          description: Storage backend type
        download_urls:
          type: object
          additionalProperties:
            $ref: '#/components/schemas/FileSignedURLDetails'
          description: Map of filename to presigned download URL details
        error_code:
          type:
          - string
          - 'null'
        error_message:
          type:
          - string
          - 'null'
      required:
      - job_id
      - job_state
      - storage_container_type
      - download_urls
      description: Response for download-files endpoint.
      title: DocDigitizationDownloadFilesResponse
    DocDigitizationUploadFilesRequest:
      type: object
      properties:
        job_id:
          type: string
          description: Job identifier returned from Create Job
        files:
          type: array
          items:
            type: string
          description: 'List of filenames to upload (exactly 1 file: PDF or ZIP)'
      required:
      - job_id
      - files
      description: Request to get upload URLs for job files
      title: DocDigitizationUploadFilesRequest
    DocDigitizationUploadFilesResponse:
      type: object
      properties:
        job_id:
          type: string
          format: uuid
          description: Job identifier
        job_state:
          $ref: '#/components/schemas/DocDigitizationJobState'
          description: Current job state
        upload_urls:
          type: object
          additionalProperties:
            $ref: '#/components/schemas/FileSignedURLDetails'
          description: Map of filename to presigned upload URL details
        storage_container_type:
          $ref: '#/components/schemas/StorageContainerType'
          description: Storage backend type
      required:
      - job_id
      - job_state
      - upload_urls
      - storage_container_type
      description: Response with presigned upload URLs
      title: DocDigitizationUploadFilesResponse
  securitySchemes:
    ApiKeyAuth:
      type: apiKey
      in: header
      name: api-subscription-key
    HTTPBearer:
      type: http
      scheme: bearer