What defines a company in Captide?
Which company identifiers does Captide use?
companyId
) to uniquely identify each entity. Ticker symbols, when paired with their respective exchanges, also serve as reliable identifiers. Users can retrieve the full list of available companies—along with their companyId
and tickers
via the /api/v1/companies/list
endpoint.When are new documents available?
Are all document types available for every company?
How do available document dates differ? Is point-in-time data available?
documentDate
: Indicates the date the document refers to (e.g., the press release date for a current report, or the fiscal period end date for annual and interim reports).filingDate
: The date and time the document was made publicly available (useful for point-in-time analysis).fiscalYear
and fiscalQuarter
: Indicate the fiscal period covered by the document.createdAt
: The date and time the document became available in Captide.updatedAt
: The most recent date and time the document was updated in Captide.What is the use case of Markdown versions of documents?
What is different between the agent-response and agent-response-stream endpoints?
agent-response
endpoint is optimized for use cases where the full response is only needed after generation is complete (e.g., data extraction or alert systems). In contrast, the agent-response-stream
endpoint streams the response in real time via server-sent events, making it more suitable for chat interfaces and interactive applications.What is the use case of the chunk-retrieval endpoint?
chunk-retrieval
endpoint is designed for firms that want to retrieve relevant content from corporate disclosures in response to natural language prompts—without involving an LLM, or where greater control over AI-generated responses is required. This is particularly useful for applications that aggregate context from multiple sources, allowing Captide’s retrieved content to be combined with data from other providers or internal systems.How can I verify the source of AI generated answers?
agent-response
and agent-response-stream
endpoints provide sentence-level attribution to excerpts from corporate disclosures. Clients can programmatically access these excerpts and metadata, or use the Captide SDK to easily display sources within a document viewer.