PostHog evaluations API

openapi: 3.0.3
info:
  title: PostHog actions evaluations API
  version: 1.0.0
  description: ''
tags:
- name: evaluations
paths:
  /api/environments/{project_id}/evaluations/:
    get:
      operationId: evaluations_list
      parameters:
      - in: query
        name: enabled
        schema:
          type: boolean
        description: Filter by enabled status
      - in: query
        name: id__in
        schema:
          type: array
          items:
            type: string
            format: uuid
        description: Multiple values may be separated by commas.
        explode: false
        style: form
      - name: limit
        required: false
        in: query
        description: Number of results to return per page.
        schema:
          type: integer
      - name: offset
        required: false
        in: query
        description: The initial index from which to return the results.
        schema:
          type: integer
      - in: query
        name: order_by
        schema:
          type: array
          items:
            type: string
            enum:
            - -created_at
            - -name
            - -updated_at
            - created_at
            - name
            - updated_at
        description: 'Ordering


          * `created_at` - Created At

          * `-created_at` - Created At (descending)

          * `updated_at` - Updated At

          * `-updated_at` - Updated At (descending)

          * `name` - Name

          * `-name` - Name (descending)'
        explode: false
        style: form
      - $ref: '#/components/parameters/ProjectIdPath'
      - in: query
        name: search
        schema:
          type: string
        description: Search in name or description
      tags:
      - evaluations
      security:
      - PersonalAPIKeyAuth:
        - evaluation:read
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/PaginatedEvaluationList'
          description: ''
      x-explicit-tags:
      - llm_analytics
    post:
      operationId: evaluations_create
      parameters:
      - $ref: '#/components/parameters/ProjectIdPath'
      tags:
      - evaluations
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/Evaluation'
          application/x-www-form-urlencoded:
            schema:
              $ref: '#/components/schemas/Evaluation'
          multipart/form-data:
            schema:
              $ref: '#/components/schemas/Evaluation'
        required: true
      security:
      - PersonalAPIKeyAuth:
        - evaluation:write
      responses:
        '201':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Evaluation'
          description: ''
      x-explicit-tags:
      - llm_analytics
  /api/environments/{project_id}/evaluations/{id}/:
    get:
      operationId: evaluations_retrieve
      parameters:
      - in: path
        name: id
        schema:
          type: string
          format: uuid
        description: A UUID string identifying this evaluation.
        required: true
      - $ref: '#/components/parameters/ProjectIdPath'
      tags:
      - evaluations
      security:
      - PersonalAPIKeyAuth:
        - evaluation:read
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Evaluation'
          description: ''
      x-explicit-tags:
      - llm_analytics
    put:
      operationId: evaluations_update
      parameters:
      - in: path
        name: id
        schema:
          type: string
          format: uuid
        description: A UUID string identifying this evaluation.
        required: true
      - $ref: '#/components/parameters/ProjectIdPath'
      tags:
      - evaluations
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/Evaluation'
          application/x-www-form-urlencoded:
            schema:
              $ref: '#/components/schemas/Evaluation'
          multipart/form-data:
            schema:
              $ref: '#/components/schemas/Evaluation'
        required: true
      security:
      - PersonalAPIKeyAuth:
        - evaluation:write
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Evaluation'
          description: ''
      x-explicit-tags:
      - llm_analytics
    patch:
      operationId: evaluations_partial_update
      parameters:
      - in: path
        name: id
        schema:
          type: string
          format: uuid
        description: A UUID string identifying this evaluation.
        required: true
      - $ref: '#/components/parameters/ProjectIdPath'
      tags:
      - evaluations
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/PatchedEvaluation'
          application/x-www-form-urlencoded:
            schema:
              $ref: '#/components/schemas/PatchedEvaluation'
          multipart/form-data:
            schema:
              $ref: '#/components/schemas/PatchedEvaluation'
      security:
      - PersonalAPIKeyAuth:
        - evaluation:write
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Evaluation'
          description: ''
      x-explicit-tags:
      - llm_analytics
    delete:
      operationId: evaluations_destroy
      description: Hard delete of this model is not allowed. Use a patch API call to set "deleted" to true
      parameters:
      - in: path
        name: id
        schema:
          type: string
          format: uuid
        description: A UUID string identifying this evaluation.
        required: true
      - $ref: '#/components/parameters/ProjectIdPath'
      tags:
      - evaluations
      security:
      - PersonalAPIKeyAuth:
        - evaluation:write
      responses:
        '405':
          description: No response body
      x-explicit-tags:
      - llm_analytics
  /api/environments/{project_id}/evaluations/test_hog/:
    post:
      operationId: evaluations_test_hog_create
      description: Test Hog evaluation code against sample events without saving.
      parameters:
      - $ref: '#/components/parameters/ProjectIdPath'
      tags:
      - evaluations
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/TestHogRequest'
          application/x-www-form-urlencoded:
            schema:
              $ref: '#/components/schemas/TestHogRequest'
          multipart/form-data:
            schema:
              $ref: '#/components/schemas/TestHogRequest'
        required: true
      security:
      - PersonalAPIKeyAuth:
        - evaluation:read
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/TestHogResponse'
          description: ''
      x-explicit-tags:
      - llm_analytics
components:
  schemas:
    TestHogResponse:
      type: object
      properties:
        results:
          type: array
          items:
            $ref: '#/components/schemas/TestHogResultItem'
        message:
          type: string
          description: Optional message, e.g. when no recent events were found.
      required:
      - results
    BlankEnum:
      enum:
      - ''
    TestHogRequest:
      type: object
      properties:
        source:
          type: string
          description: Hog source code to test. Must return a boolean (true = pass, false = fail) or null for N/A.
          minLength: 1
        sample_count:
          type: integer
          maximum: 10
          minimum: 1
          default: 5
          description: Number of recent $ai_generation events to test against (1–10, default 5).
        allows_na:
          type: boolean
          default: false
          description: Whether the evaluation can return N/A for non-applicable generations.
        conditions:
          type: array
          items:
            type: object
            additionalProperties: true
          description: Optional trigger conditions to filter which events are sampled.
      required:
      - source
    TestHogResultItem:
      type: object
      properties:
        event_uuid:
          type: string
          description: UUID of the $ai_generation event.
        trace_id:
          type: string
          nullable: true
          description: Trace ID if available.
        input_preview:
          type: string
          description: First 200 chars of the generation input.
        output_preview:
          type: string
          description: First 200 chars of the generation output.
        result:
          type: boolean
          nullable: true
          description: True = pass, False = fail, null = N/A or error.
        reasoning:
          type: string
          nullable: true
          description: Hog evaluation reasoning string, if any.
        error:
          type: string
          nullable: true
          description: Error message if the Hog code raised an exception.
      required:
      - error
      - event_uuid
      - input_preview
      - output_preview
      - reasoning
      - result
    LLMProviderEnum:
      enum:
      - openai
      - anthropic
      - gemini
      - openrouter
      - fireworks
      - azure_openai
      - together_ai
      type: string
      description: '* `openai` - Openai

        * `anthropic` - Anthropic

        * `gemini` - Gemini

        * `openrouter` - Openrouter

        * `fireworks` - Fireworks

        * `azure_openai` - Azure OpenAI

        * `together_ai` - Together AI'
    UserBasic:
      type: object
      properties:
        id:
          type: integer
          readOnly: true
        uuid:
          type: string
          format: uuid
          readOnly: true
        distinct_id:
          type: string
          nullable: true
          maxLength: 200
        first_name:
          type: string
          maxLength: 150
        last_name:
          type: string
          maxLength: 150
        email:
          type: string
          format: email
          title: Email address
          maxLength: 254
        is_email_verified:
          type: boolean
          nullable: true
        hedgehog_config:
          type: object
          additionalProperties: true
          nullable: true
          readOnly: true
        role_at_organization:
          nullable: true
          oneOf:
          - $ref: '#/components/schemas/RoleAtOrganizationEnum'
          - $ref: '#/components/schemas/BlankEnum'
          - $ref: '#/components/schemas/NullEnum'
      required:
      - email
      - hedgehog_config
      - id
      - uuid
    EvaluationTypeEnum:
      enum:
      - llm_judge
      - hog
      type: string
      description: '* `llm_judge` - LLM as a judge

        * `hog` - Hog'
    StatusReasonEnum:
      enum:
      - trial_limit_reached
      - model_not_allowed
      - provider_key_deleted
      type: string
      description: '* `trial_limit_reached` - Trial evaluation limit reached

        * `model_not_allowed` - Model not available on the trial plan

        * `provider_key_deleted` - Provider API key was deleted'
    Evaluation:
      type: object
      properties:
        id:
          type: string
          format: uuid
          readOnly: true
        name:
          type: string
          description: Name of the evaluation.
          maxLength: 400
        description:
          type: string
          description: Optional description of what this evaluation checks.
        enabled:
          type: boolean
          description: Whether the evaluation runs automatically on new $ai_generation events.
        status:
          allOf:
          - $ref: '#/components/schemas/EvaluationStatusEnum'
          readOnly: true
        status_reason:
          readOnly: true
          nullable: true
          oneOf:
          - $ref: '#/components/schemas/StatusReasonEnum'
          - $ref: '#/components/schemas/NullEnum'
        evaluation_type:
          allOf:
          - $ref: '#/components/schemas/EvaluationTypeEnum'
          description: '''llm_judge'' uses an LLM to score outputs against a prompt; ''hog'' runs deterministic Hog code.


            * `llm_judge` - LLM as a judge

            * `hog` - Hog'
        evaluation_config:
          oneOf:
          - type: object
            title: LLM judge config
            required:
            - prompt
            properties:
              prompt:
                type: string
                description: Evaluation criteria for the LLM judge. Describe what makes a good vs bad response.
                minLength: 1
            additionalProperties: false
          - type: object
            title: Hog config
            required:
            - source
            properties:
              source:
                type: string
                description: Hog source code. Must return true (pass), false (fail), or null for N/A.
                minLength: 1
            additionalProperties: false
          description: 'Configuration dict. For ''llm_judge'': {prompt}. For ''hog'': {source}.'
        output_type:
          allOf:
          - $ref: '#/components/schemas/OutputTypeEnum'
          description: 'Output format. Currently only ''boolean'' is supported.


            * `boolean` - Boolean (Pass/Fail)'
        output_config:
          type: object
          properties:
            allows_na:
              type: boolean
              description: Whether the evaluation can return N/A for non-applicable generations.
              default: false
          additionalProperties: false
          description: 'Output config. For ''boolean'' output_type: {allows_na} to permit N/A results.'
        conditions:
          description: Optional trigger conditions to filter which events are evaluated. OR between condition sets, AND within each.
        model_configuration:
          allOf:
          - $ref: '#/components/schemas/ModelConfiguration'
          nullable: true
        created_at:
          type: string
          format: date-time
          readOnly: true
        updated_at:
          type: string
          format: date-time
          readOnly: true
        created_by:
          allOf:
          - $ref: '#/components/schemas/UserBasic'
          readOnly: true
        deleted:
          type: boolean
          description: Set to true to soft-delete the evaluation.
      required:
      - created_at
      - created_by
      - evaluation_type
      - id
      - name
      - output_type
      - status
      - status_reason
      - updated_at
    RoleAtOrganizationEnum:
      enum:
      - engineering
      - data
      - product
      - founder
      - leadership
      - marketing
      - sales
      - other
      type: string
      description: '* `engineering` - Engineering

        * `data` - Data

        * `product` - Product Management

        * `founder` - Founder

        * `leadership` - Leadership

        * `marketing` - Marketing

        * `sales` - Sales / Success

        * `other` - Other'
    EvaluationStatusEnum:
      enum:
      - active
      - paused
      - error
      type: string
      description: '* `active` - Active

        * `paused` - Paused

        * `error` - Error'
    ModelConfiguration:
      type: object
      description: Nested serializer for model configuration.
      properties:
        provider:
          $ref: '#/components/schemas/LLMProviderEnum'
        model:
          type: string
          maxLength: 100
        provider_key_id:
          type: string
          format: uuid
          nullable: true
        provider_key_name:
          type: string
          nullable: true
          readOnly: true
      required:
      - model
      - provider
      - provider_key_name
    OutputTypeEnum:
      enum:
      - boolean
      type: string
      description: '* `boolean` - Boolean (Pass/Fail)'
    PatchedEvaluation:
      type: object
      properties:
        id:
          type: string
          format: uuid
          readOnly: true
        name:
          type: string
          description: Name of the evaluation.
          maxLength: 400
        description:
          type: string
          description: Optional description of what this evaluation checks.
        enabled:
          type: boolean
          description: Whether the evaluation runs automatically on new $ai_generation events.
        status:
          allOf:
          - $ref: '#/components/schemas/EvaluationStatusEnum'
          readOnly: true
        status_reason:
          readOnly: true
          nullable: true
          oneOf:
          - $ref: '#/components/schemas/StatusReasonEnum'
          - $ref: '#/components/schemas/NullEnum'
        evaluation_type:
          allOf:
          - $ref: '#/components/schemas/EvaluationTypeEnum'
          description: '''llm_judge'' uses an LLM to score outputs against a prompt; ''hog'' runs deterministic Hog code.


            * `llm_judge` - LLM as a judge

            * `hog` - Hog'
        evaluation_config:
          oneOf:
          - type: object
            title: LLM judge config
            required:
            - prompt
            properties:
              prompt:
                type: string
                description: Evaluation criteria for the LLM judge. Describe what makes a good vs bad response.
                minLength: 1
            additionalProperties: false
          - type: object
            title: Hog config
            required:
            - source
            properties:
              source:
                type: string
                description: Hog source code. Must return true (pass), false (fail), or null for N/A.
                minLength: 1
            additionalProperties: false
          description: 'Configuration dict. For ''llm_judge'': {prompt}. For ''hog'': {source}.'
        output_type:
          allOf:
          - $ref: '#/components/schemas/OutputTypeEnum'
          description: 'Output format. Currently only ''boolean'' is supported.


            * `boolean` - Boolean (Pass/Fail)'
        output_config:
          type: object
          properties:
            allows_na:
              type: boolean
              description: Whether the evaluation can return N/A for non-applicable generations.
              default: false
          additionalProperties: false
          description: 'Output config. For ''boolean'' output_type: {allows_na} to permit N/A results.'
        conditions:
          description: Optional trigger conditions to filter which events are evaluated. OR between condition sets, AND within each.
        model_configuration:
          allOf:
          - $ref: '#/components/schemas/ModelConfiguration'
          nullable: true
        created_at:
          type: string
          format: date-time
          readOnly: true
        updated_at:
          type: string
          format: date-time
          readOnly: true
        created_by:
          allOf:
          - $ref: '#/components/schemas/UserBasic'
          readOnly: true
        deleted:
          type: boolean
          description: Set to true to soft-delete the evaluation.
    PaginatedEvaluationList:
      type: object
      required:
      - count
      - results
      properties:
        count:
          type: integer
          example: 123
        next:
          type: string
          nullable: true
          format: uri
          example: http://api.example.org/accounts/?offset=400&limit=100
        previous:
          type: string
          nullable: true
          format: uri
          example: http://api.example.org/accounts/?offset=200&limit=100
        results:
          type: array
          items:
            $ref: '#/components/schemas/Evaluation'
    NullEnum:
      enum:
      - null
  parameters:
    ProjectIdPath:
      in: path
      name: project_id
      required: true
      schema:
        type: string
      description: Project ID of the project you're trying to access. To find the ID of the project, make a call to /api/projects/.
  securitySchemes:
    PersonalAPIKeyAuth:
      type: http
      scheme: bearer
x-tagGroups:
- name: All endpoints
  tags:
  - LLM Analytics
  - actions
  - activity_log
  - activity_logs
  - advanced_activity_logs
  - alerts
  - annotations
  - approval_policies
  - batch_exports
  - cdp
  - change_requests
  - code
  - code-invites
  - cohorts
  - comments
  - conversations
  - core
  - customer_analytics
  - customer_journeys
  - customer_profile_configs
  - dashboard_templates
  - dashboards
  - data_color_themes
  - data_modeling_jobs
  - data_warehouse
  - dataset_items
  - datasets
  - desktop_recordings
  - domains
  - early_access_feature
  - early_access_features
  - elements
  - endpoints
  - environments
  - error_tracking
  - evaluation_runs
  - evaluations
  - event_definitions
  - event_filter
  - event_schemas
  - events
  - experiment_holdouts
  - experiment_saved_metrics
  - experiments
  - exports
  - external_data_schemas
  - external_data_sources
  - feature_flags
  - file_system
  - file_system_shortcut
  - flag_value
  - groups
  - groups_types
  - health_issues
  - heatmap_screenshots
  - heatmaps
  - hog_flows
  - hog_function_templates
  - hog_functions
  - insight_variables
  - insights
  - integrations
  - invites
  - js-snippet
  - legal_documents
  - lineage
  - live_debugger_breakpoints
  - llm_analytics
  - llm_prompts
  - llm_skills
  - logs
  - managed_viewsets
  - max
  - max_tools
  - mcp_server_installations
  - mcp_servers
  - mcp_store
  - mcp_tools
  - members
  - notebooks
  - oauth_applications
  - object_media_previews
  - organizations
  - persisted_folder
  - persons
  - platform_features
  - plugin_configs
  - product_analytics
  - product_tours
  - project_secret_api_keys
  - projects
  - property_definitions
  - proxy_records
  - public_hog_function_templates
  - query
  - replay
  - reverse_proxy
  - role_external_references
  - roles
  - sandbox-environments
  - sandbox_environments
  - saved
  - schema_property_groups
  - sdk_doctor
  - session_group_summaries
  - session_recording_playlists
  - session_recordings
  - session_summaries
  - sessions
  - signals
  - subscriptions
  - surveys
  - taggers
  - task-automations
  - task-runs
  - task_automations
  - tasks
  - uploaded_media
  - user_home_settings
  - user_interviews
  - users
  - visual_review
  - warehouse_dag
  - warehouse_model_paths
  - warehouse_saved_queries
  - warehouse_saved_query_folders
  - warehouse_tables
  - warehouse_view_link
  - warehouse_view_links
  - web_analytics
  - web_experiments
  - web_vitals
  - welcome
  - workflows