> ## Documentation Index
> Fetch the complete documentation index at: https://docs-platform.crewai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Get an Automation's eval scores

> Returns the points of the Evals tab's charts: per criterion, its average score in each time bucket, one row
per deployment version, as the charts draw a line per version. Buckets are UTC hours for `24h` and UTC days
for `7d` and `30d`. With `at`, returns the one bucket that holds that moment.

Criteria are listed as the criteria endpoint lists them for the same range; a criterion nothing was graded on
has no rows.




## OpenAPI

````yaml /openapi/platform-v1.yaml get /api/v1/automations/{automation_id}/evals/scores
openapi: 3.0.1
info:
  title: CrewAI Platform API
  version: v1
  description: Supported public API for CrewAI Platform.
servers:
  - url: https://app.crewai.com
    description: Current CrewAI Platform host
security:
  - bearerAuth: []
tags:
  - name: Users
    description: Users in the organization.
  - name: Status
    description: Platform health and API availability.
  - name: Automations
    description: >-
      Automations, their source artifacts, their executions, and their
      schedules.
  - name: Teams
    description: Teams in the organization.
  - name: Group Mappings
    description: Identity-provider group mappings to Teams.
  - name: Service Accounts
    description: Service accounts in the organization.
  - name: Webhook Endpoints
    description: Endpoints that receive signed event deliveries.
  - name: Studio Projects
    description: Projects built in Crew Studio.
  - name: Environment Variables
    description: Environment variables in the organization.
  - name: Git Repositories
    description: Git repositories connected to the organization.
  - name: LLM Connections
    description: Connections to LLM providers.
  - name: Tools
    description: Tools in the organization's tool repository.
  - name: Applications
    description: Applications the organization can connect, with their tools and triggers.
  - name: Connections
    description: Application connections the caller can use.
  - name: Evals
    description: How an Automation's runs score on its continuous evals criteria over time.
paths:
  /api/v1/automations/{automation_id}/evals/scores:
    parameters:
      - name: automation_id
        in: path
        required: true
        schema:
          type: string
          format: uuid
          example: 4f6d0d9a-7b78-4c6a-b64f-2e2e3d2c8f9a
      - name: range
        in: query
        required: false
        schema:
          type: string
          enum:
            - 24h
            - 7d
            - 30d
          default: 24h
          x-crewai-enum-format: freeform
          description: >
            The time range, ending now, as the Evals tab's range picker offers
            it:

            - `24h`: The last 24 hours, in hourly buckets.

            - `7d`: The last 7 days, in daily buckets.

            - `30d`: The last 30 days, in daily buckets.
      - name: at
        in: query
        required: false
        description: >-
          A moment instead of a range: a date (`2026-11-18`) returns that UTC
          day's daily bucket, an RFC 3339 date-time with a UTC offset
          (`2026-11-18T14:00:00Z`) the hourly bucket it falls in. Pass `range`
          or `at`, not both.
        schema:
          type: string
          example: '2026-11-18'
      - name: criteria
        in: query
        required: false
        description: >-
          `all`, or a comma-separated list of criterion keys from the criteria
          endpoint. A key that the criteria endpoint does not list for the same
          range or `at` is `not_found`.
        schema:
          type: string
          default: all
          example: crew:brief,task:publish:7d59a677c485
    get:
      tags:
        - Evals
      summary: Get an Automation's eval scores
      description: >
        Returns the points of the Evals tab's charts: per criterion, its average
        score in each time bucket, one row

        per deployment version, as the charts draw a line per version. Buckets
        are UTC hours for `24h` and UTC days

        for `7d` and `30d`. With `at`, returns the one bucket that holds that
        moment.


        Criteria are listed as the criteria endpoint lists them for the same
        range; a criterion nothing was graded on

        has no rows.
      operationId: getAutomationEvalScores
      responses:
        '200':
          description: OK
          content:
            application/json:
              examples:
                success:
                  value:
                    data:
                      - key: crew:brief
                        rows:
                          - time: '2026-11-18T10:00:00Z'
                            deployment_version: 1
                            score: 2.5
                            samples: 1
                          - time: '2026-11-18T13:00:00Z'
                            deployment_version: 2
                            score: 7.5
                            samples: 2
                    meta:
                      bucket: hour
                      start_time: '2026-11-17T15:30:00Z'
                      end_time: '2026-11-18T15:30:00Z'
              schema:
                type: object
                required:
                  - data
                  - meta
                additionalProperties: false
                properties:
                  data:
                    type: array
                    items:
                      type: object
                      required:
                        - key
                        - rows
                      additionalProperties: false
                      properties:
                        key:
                          type: string
                          example: crew:brief
                          description: The criterion's key.
                        rows:
                          type: array
                          description: >-
                            One row per time bucket and deployment version that
                            has graded runs, ordered by `time`, then
                            `deployment_version`. Buckets with nothing graded
                            are left out.
                          items:
                            type: object
                            required:
                              - time
                              - deployment_version
                              - score
                              - samples
                            additionalProperties: false
                            properties:
                              time:
                                type: string
                                format: date-time
                                example: '2026-11-18T13:00:00Z'
                                description: Where the bucket starts, in UTC.
                              deployment_version:
                                type: integer
                                minimum: 1
                                example: 2
                                description: >-
                                  The deployment version the graded runs ran on:
                                  n in the Evals tab's "Deployment #n", counting
                                  the Automation's deployments from its first.
                              score:
                                type: number
                                minimum: 0
                                maximum: 10
                                example: 7.5
                                description: >-
                                  The average score from 0 to 10 of the runs
                                  graded in the bucket: the share of the
                                  criterion's checks that passed, times 10.
                              samples:
                                type: integer
                                minimum: 1
                                example: 2
                                description: How many graded runs the score averages.
                  meta:
                    type: object
                    required:
                      - bucket
                      - start_time
                      - end_time
                    additionalProperties: false
                    description: The window the response covers.
                    properties:
                      bucket:
                        type: string
                        enum:
                          - hour
                          - day
                        example: hour
                        description: >
                          How long each score's time bucket is. Buckets are UTC
                          and start on the hour or at midnight:

                          - `hour`: One hour.

                          - `day`: One UTC day.
                      start_time:
                        type: string
                        format: date-time
                        example: '2026-11-17T15:30:00Z'
                        description: Where the window starts.
                      end_time:
                        type: string
                        format: date-time
                        example: '2026-11-18T15:30:00Z'
                        description: >-
                          Where the window ends: now for a `range`, or the end
                          of the bucket `at` names (exclusive).
        '400':
          description: Bad Request
          content:
            application/json:
              examples:
                bad_request:
                  value:
                    errors:
                      - type: >-
                          https://docs-platform.crewai.com/api/problems/bad_request
                        code: bad_request
                        title: Bad request
                        status: 400
                        detail: >-
                          at must be a date (2026-11-18) or an RFC 3339
                          date-time with a UTC offset (2026-11-18T14:00:00Z).
                  summary: An `at` that is not a date or a date-time
              schema:
                $ref: '#/components/schemas/ErrorEnvelope'
        '401':
          description: Unauthorized
          content:
            application/json:
              examples:
                unauthorized:
                  value:
                    errors:
                      - type: >-
                          https://docs-platform.crewai.com/api/problems/unauthorized
                        code: unauthorized
                        title: Unauthorized
                        status: 401
                        detail: Missing or invalid bearer token.
                  summary: Missing or invalid bearer token
              schema:
                $ref: '#/components/schemas/ErrorEnvelope'
        '404':
          description: Not Found
          content:
            application/json:
              examples:
                not_found:
                  value:
                    errors:
                      - type: >-
                          https://docs-platform.crewai.com/api/problems/not_found
                        code: not_found
                        title: Not found
                        status: 404
                        detail: >-
                          No criterion in this range has the key
                          "crew:no-such-criterion".
                  summary: A criterion the range has no chart for
              schema:
                $ref: '#/components/schemas/ErrorEnvelope'
components:
  schemas:
    ErrorEnvelope:
      type: object
      required:
        - errors
      additionalProperties: false
      properties:
        errors:
          type: array
          minItems: 1
          items:
            $ref: '#/components/schemas/Error'
    Error:
      type: object
      description: Public API error object.
      required:
        - type
        - code
        - title
        - status
        - detail
      additionalProperties: false
      properties:
        type:
          type: string
          format: uri
          enum:
            - https://docs-platform.crewai.com/api/problems/bad_request
            - https://docs-platform.crewai.com/api/problems/clipper_disabled
            - https://docs-platform.crewai.com/api/problems/clipper_disallowed
            - >-
              https://docs-platform.crewai.com/api/problems/clipper_feature_disabled
            - https://docs-platform.crewai.com/api/problems/clipper_not_found
            - https://docs-platform.crewai.com/api/problems/conflict
            - https://docs-platform.crewai.com/api/problems/content_too_large
            - https://docs-platform.crewai.com/api/problems/forbidden
            - https://docs-platform.crewai.com/api/problems/internal_error
            - https://docs-platform.crewai.com/api/problems/not_found
            - https://docs-platform.crewai.com/api/problems/service_unavailable
            - >-
              https://docs-platform.crewai.com/api/problems/tool_execution_failed
            - https://docs-platform.crewai.com/api/problems/unauthorized
            - https://docs-platform.crewai.com/api/problems/validation_error
          example: https://docs-platform.crewai.com/api/problems/bad_request
        code:
          type: string
          enum:
            - bad_request
            - clipper_disabled
            - clipper_disallowed
            - clipper_feature_disabled
            - clipper_not_found
            - conflict
            - content_too_large
            - forbidden
            - internal_error
            - not_found
            - service_unavailable
            - tool_execution_failed
            - unauthorized
            - validation_error
          example: bad_request
        title:
          type: string
          x-crewai-enum-format: freeform
          enum:
            - Bad request
            - Clipper resource disabled
            - Clipper resource disallowed
            - Clipper feature disabled
            - Clipper resource not found
            - Conflict
            - Content too large
            - Forbidden
            - Internal error
            - Not found
            - Service unavailable
            - Tool execution failed
            - Unauthorized
            - Validation error
          example: Bad request
        status:
          type: integer
          enum:
            - 400
            - 401
            - 403
            - 404
            - 409
            - 413
            - 422
            - 500
            - 503
          example: 400
        detail:
          type: string
          example: The request is invalid.
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      bearerFormat: opaque

````

This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.