Create dataset from CSV
Creates a new dataset by uploading a CSV file. Each row in the CSV becomes an entity. Each row requires a name plus at least one of: a description or a domain; all other columns are optional. Note: description in the CSV is the entity’s matching description — it is separate from the dataset-level description field in the form data.
CSV format:
name,description,domain,alternative_names,key_persons
NewsCatcher,"NewsCatcher is a data-as-a-service company providing news intelligence APIs including the CatchAll Web Search API (2B+ web pages indexed) and News API (140,000+ sources, 100+ countries).",newscatcherapi.com,"NewsCatcher CatchAll;NewsCatcher API","Artem Bugara;Maksym Sugonyaka"
OpenAI,"Artificial intelligence research company",openai.com,"Open AI","Sam Altman"
Use semicolons (;) to separate multiple values in alternative_names and key_persons. Rows with empty name are skipped and reported in validation_report.
Note: The response shape differs from the JSON dataset creation endpoint: it returns dataset_id (not id) and includes a validation_report with details on skipped rows.
Authorizations
API key for authentication.
Body
Response
Dataset created from CSV successfully.
Unique identifier of the created dataset.
"ccabb755-afc2-4047-b84c-78d1f23d49b2"
Name of the created dataset.
"My Portfolio"
Number of entities successfully created from the CSV.
3
Summary of CSV processing results.

