> ## Documentation Index > Fetch the complete documentation index at: https://docs.nx1cloud.com/llms.txt > Use this file to discover all available pages before exploring further. # Create a fine-tuning dataset proposal > Start a CrewAI run that assembles domain knowledge into Q/A examples. Crew events are published under the job's UUID; the portal polls ``/api/events/{job_id}``. The response carries that ID as ``X-Correlation-ID``. ## OpenAPI ````yaml post /api/finetune/datasets openapi: 3.1.0 info: title: Nx1 AI API description: | AI API for Nx1 Data Platform Management and Automated Data Tasks. Authentication is required via PSK in Authorization header. Default PSK is | [ask a friend] | version: 0.10.2 servers: [] security: [] paths: /api/finetune/datasets: post: tags: - AutoLLM summary: Create a fine-tuning dataset proposal description: >- Start a CrewAI run that assembles domain knowledge into Q/A examples. Crew events are published under the job's UUID; the portal polls ``/api/events/{job_id}``. The response carries that ID as ``X-Correlation-ID``. operationId: create_dataset_api_finetune_datasets_post requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/FineTuneDatasetRequest' responses: '202': description: Crew run started. Poll the job for the proposal. content: application/json: schema: $ref: '#/components/schemas/FineTuneDatasetProposal' '400': content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' description: Bad Request '422': description: Validation Error content: application/json: schema: $ref: '#/components/schemas/HTTPValidationError' '500': content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' description: Internal Server Error security: - OAuth2AuthorizationCodeBearer: [] - APIKeyHeader: [] components: schemas: FineTuneDatasetRequest: properties: domain: type: string maxLength: 512 title: Domain description: DataHub domain name or URN whose knowledge seeds the dataset. base_model: type: string maxLength: 256 title: Base Model description: >- Base model this dataset will fine-tune. Must be present in FINETUNE_BASE_MODEL_ALLOWLIST; validated server-side. name: anyOf: - type: string maxLength: 256 - type: 'null' title: Name description: Human-friendly dataset name. Defaults to '-finetune'. extra_instructions: anyOf: - type: string maxLength: 20000 - type: 'null' title: Extra Instructions description: >- Free-form guidance and facts the user wants baked in — extra context the crew should weave into the training examples. target_examples: anyOf: - type: integer maximum: 500 minimum: 1 - type: 'null' title: Target Examples description: >- Soft target for how many Q/A examples to generate. Clamped by FINETUNE_DATASET_CREW_MAX_EXAMPLES. include_glossary: type: boolean title: Include Glossary default: true include_documents: type: boolean title: Include Documents default: true include_lineage: type: boolean title: Include Lineage default: true include_tags: type: boolean title: Include Tags default: true include_data_products: type: boolean title: Include Data Products default: true owner: anyOf: - type: string maxLength: 256 - type: 'null' title: Owner description: >- Owner username. Ignored for non-admin callers (the authenticated user is always the owner); admins may set it to create on behalf of another user. type: object required: - domain - base_model title: FineTuneDatasetRequest description: Start a dataset-generation run for a DataHub domain. FineTuneDatasetProposal: properties: job_id: type: string format: uuid title: Job Id status: $ref: '#/components/schemas/FineTuneDatasetStatusEnum' name: type: string title: Name domain: type: string title: Domain base_model: type: string title: Base Model owner: type: string title: Owner example_count: type: integer title: Example Count default: 0 coverage: anyOf: - $ref: '#/components/schemas/CoverageStats' - type: 'null' sample_examples: items: $ref: '#/components/schemas/QAExample' type: array title: Sample Examples notes: anyOf: - type: string - type: 'null' title: Notes proposed_dataset_s3: anyOf: - type: string - type: 'null' title: Proposed Dataset S3 approved_dataset_s3: anyOf: - type: string - type: 'null' title: Approved Dataset S3 mlflow_dataset_uri: anyOf: - type: string - type: 'null' title: Mlflow Dataset Uri error: anyOf: - type: string - type: 'null' title: Error created_at: type: string format: date-time title: Created At updated_at: type: string format: date-time title: Updated At type: object required: - job_id - status - name - domain - base_model - owner - created_at - updated_at title: FineTuneDatasetProposal description: >- The crew's proposed dataset, returned for review. The full example set lives in S3 (``proposed_dataset_s3`` / after approval, ``approved_dataset_s3``). ``sample_examples`` is a bounded preview. ErrorResponse: properties: error: type: string title: Error description: A brief description of the error that occurred. code: type: integer title: Code description: The HTTP status code associated with the error. default: 500 type: object required: - error title: ErrorResponse HTTPValidationError: properties: detail: items: $ref: '#/components/schemas/ValidationError' type: array title: Detail type: object title: HTTPValidationError FineTuneDatasetStatusEnum: type: string enum: - running - pending_approval - failed - approved - materializing - complete title: FineTuneDatasetStatusEnum description: Lifecycle of a fine-tuning dataset job. CoverageStats: properties: glossary_terms: type: integer title: Glossary Terms default: 0 documents: type: integer title: Documents default: 0 datasets: type: integer title: Datasets default: 0 tags: type: integer title: Tags default: 0 data_products: type: integer title: Data Products default: 0 total_examples: type: integer title: Total Examples default: 0 type: object title: CoverageStats description: How much of the domain the crew drew on. QAExample: properties: messages: items: $ref: '#/components/schemas/ChatMessage' type: array maxItems: 50 minItems: 1 title: Messages description: Chat turns; typically a user question and an assistant answer. provenance: anyOf: - type: string - type: 'null' title: Provenance description: Source of the example — a DataHub URN, document name, or note. category: anyOf: - type: string - type: 'null' title: Category description: >- Knowledge category: glossary | document | lineage | tag | data_product | user. type: object required: - messages title: QAExample description: A single instruction-tuning example in chat format. ValidationError: properties: loc: items: anyOf: - type: string - type: integer type: array title: Location msg: type: string title: Message type: type: string title: Error Type input: title: Input ctx: type: object title: Context type: object required: - loc - msg - type title: ValidationError ChatMessage: properties: role: type: string enum: - system - user - assistant title: Role content: type: string maxLength: 100000 title: Content type: object required: - role - content title: ChatMessage description: One turn in a chat-format training example. securitySchemes: OAuth2AuthorizationCodeBearer: type: oauth2 flows: authorizationCode: scopes: {} authorizationUrl: >- https://sso-rapid.rapid.nx1cloud.com/realms/rapid/protocol/openid-connect/auth tokenUrl: >- https://sso-rapid.rapid.nx1cloud.com/realms/rapid/protocol/openid-connect/token APIKeyHeader: type: apiKey in: header name: Authorization-PSK ````