> ## Documentation Index > Fetch the complete documentation index at: https://docs.dataframer.ai/llms.txt > Use this file to discover all available pages before exploring further. # Create seed dataset > Create a new seed dataset with uploaded files You can upload one of the following types of datasets: * **SINGLE\_FILE**: exactly one file containing all the samples * **MULTI\_FILE**: multiple files, where each file is a separate sample * **MULTI\_FOLDER**: multiple folders, where each folder is a separate sample and only contains files with no nested folders ## File constraints by dataset type | Type | Allowed formats | Size limit | Count limit | | ------------- | ------------------------------ | -------------------- | -------------------------------------- | | SINGLE\_FILE | CSV, JSON, JSONL | 50MB | 1 file, min 2 rows | | MULTI\_FILE | TXT, MD, JSON, CSV, JSONL, PDF | 1MB/file, 50MB total | 2-1000 files | | MULTI\_FOLDER | TXT, MD, JSON, CSV, JSONL, PDF | 1MB/file, 50MB total | 2-1000 files, 20/folder, min 2 folders | ## OpenAPI ````yaml POST /api/dataframer/seed-datasets/create/ openapi: 3.0.0 info: title: DataFramer API version: 0.1.0 description: '' termsOfService: https://www.aimon.ai/docs/privacy-policy.pdf contact: name: DataFramer Support email: info@dataframer.ai license: name: Proprietary x-logo: url: https://dataframer.ai/logo.png altText: DataFramer AI x-stainless: package-name: aimon-dataframer namespace: - aimon - dataframer servers: - url: https://df-api.dataframer.ai description: Production server security: - BearerAuth: [] tags: - name: Seed Datasets description: Manage seed datasets for generation - name: Specs description: Data specifications for sample generation - name: Runs description: Generation runs and results - name: Evaluations description: Evaluate generated sample quality - name: Red Teaming description: Security testing and adversarial prompts - name: Spec Creation description: Create specs from datasets or from scratch (seedless) - name: Generation description: Synthetic data generation - name: API Keys description: API key management and rotation - name: Health description: Health check endpoints - name: Models description: Available AI models externalDocs: description: Complete API Guide url: https://docs.dataframer.ai/dataframer paths: /api/dataframer/seed-datasets/create/: post: tags: - Seed Datasets summary: Create seed dataset description: >- Create a new seed dataset the platform learns from, with uploaded files. After creation, use create_spec with the returned dataset ID. You can upload one of the following types of datasets: - **SINGLE_FILE**: exactly one file containing all the samples - **MULTI_FILE**: multiple files, where each file is a separate sample - **MULTI_FOLDER**: multiple folders, where each folder is a separate sample and only contains files with no nested folders ## File constraints by dataset type | Type | Allowed formats | Size limit | Count limit | |------|----------------|------------|-------------| | SINGLE_FILE | CSV, JSON, JSONL | 50MB | 1 file, min 2 rows | | MULTI_FILE | TXT, MD, JSON, CSV, JSONL, PDF | 1MB/file, 50MB total | 2-1000 files | | MULTI_FOLDER | TXT, MD, JSON, CSV, JSONL, PDF | 1MB/file, 50MB total | 2-1000 files, 20/folder, min 2 folders | operationId: api_dataframer_datasets_create_create requestBody: required: true content: multipart/form-data: schema: type: object required: - name - dataset_type - files properties: name: type: string description: Dataset name (must be unique) description: type: string description: Optional dataset description dataset_type: type: string enum: - SINGLE_FILE - MULTI_FILE - MULTI_FOLDER files: type: array items: type: string format: binary description: >- Files to upload. SINGLE_FILE: exactly 1 file. MULTI_FILE: 2+ files. MULTI_FOLDER: 2+ files with corresponding folder_names. minItems: 2 folder_names: type: array items: type: string description: >- Folder names for MULTI_FOLDER datasets. This is a parallel array with files: folder_names[i] specifies which folder files[i] belongs to (e.g., if files=['a.txt', 'b.txt'] and folder_names=['doc1', 'doc2'], then a.txt goes in doc1, b.txt goes in doc2). Minimum 2 unique folder names required. For MULTI_FOLDER, file order within each folder is preserved as uploaded — earlier files should be the ones that later files may depend on. minItems: 2 encoding: files: style: form explode: true contentType: application/octet-stream folder_names: style: form explode: true responses: '201': description: Seed dataset created successfully content: application/json: schema: $ref: '#/components/schemas/DataframerDataset' examples: dataset_created: value: id: 550e8400-e29b-41d4-a716-446655440000 name: Customer Reviews 2025 description: Product reviews with files dataset_type: SINGLE_FILE file_count: 1 folder_count: 0 created_at: '2025-01-15T10:30:00Z' updated_at: '2025-01-15T10:30:00Z' created_by_email: sarah.chen@acme.com '400': description: Bad Request - Invalid data, missing files, or validation errors '401': description: Unauthorized x-codeSamples: - lang: JavaScript source: >- import Dataframer from 'dataframer'; const client = new Dataframer({ apiKey: process.env['DATAFRAMER_API_KEY'], // This is the default and can be omitted }); const response = await client.dataframer.seedDatasets.createWithFiles({ dataset_type: 'SINGLE_FILE', files: [fs.createReadStream('path/to/file'), fs.createReadStream('path/to/file')], name: 'name', }); console.log(response.id); - lang: Python source: |- import os from dataframer import Dataframer client = Dataframer( api_key=os.environ.get("DATAFRAMER_API_KEY"), # This is the default and can be omitted ) response = client.dataframer.seed_datasets.create_with_files( dataset_type="SINGLE_FILE", files=[b"Example data", b"Example data"], name="name", ) print(response.id) components: schemas: DataframerDataset: type: object properties: id: type: string format: uuid readOnly: true description: Unique identifier for the dataset name: type: string description: Dataset name description: type: string nullable: true description: Optional description of the dataset contents or purpose dataset_type: type: string enum: - SINGLE_FILE - MULTI_FILE - MULTI_FOLDER description: >- Type of dataset structure. SINGLE_FILE: one CSV/JSON/JSONL file with tabular data. MULTI_FILE: multiple individual text files. MULTI_FOLDER: files organized into folders where each folder represents one sample. created_at: type: string format: date-time readOnly: true description: Timestamp when the dataset was created updated_at: type: string format: date-time readOnly: true description: Timestamp when the dataset was last modified created_by_email: type: string readOnly: true description: Email address of the user who created the dataset files: type: array items: $ref: '#/components/schemas/File' readOnly: true description: List of all files in the dataset folder_count: type: integer readOnly: true description: Total number of folders in the dataset file_count: type: integer readOnly: true description: Total number of files in the dataset sample_count: type: integer nullable: true readOnly: true description: >- Number of data samples in the dataset. Only populated for SINGLE_FILE datasets (e.g. number of rows in a CSV). File: type: object properties: id: type: string format: uuid readOnly: true description: Unique identifier for the file file_type: type: string enum: - json - jsonl - csv - md - txt - pdf description: >- File format. json: single JSON object or array. jsonl: newline-delimited JSON records. csv: comma-separated values. md: Markdown text. txt: plain text. pdf: PDF document. size_bytes: type: integer nullable: true description: File size in bytes sha256: type: string nullable: true description: SHA-256 hash of the file contents for integrity verification path: type: string readOnly: true description: >- Full path of the file. For files in folders, includes folder name (e.g., 'folder_name/file.txt'). For files at root level, just the filename. securitySchemes: BearerAuth: type: http scheme: bearer bearerFormat: API Key description: 'API Key authentication. Format: "Bearer YOUR_API_KEY"' ````