Research data that keeps its evidence
DataCedar connects historical market observations to the filings, facts, earnings events, transcripts, and public news references that explain what investors could know at the time.
Why we built it
Most stock research begins by reconciling several providers with different identifiers, timestamps, revisions, gaps, and permissions. A price series can look complete while excluding delisted securities. A current earnings calendar can overwrite the dates expected historically. A normalized fundamental can lose the accession and XBRL concept needed to verify it.
DataCedar gives those streams one company identity and a two-clock model: when something was economically effective and when it became publicly observable. Coverage, source, and rights remain visible beside the data instead of disappearing behind an endpoint.
How the system works
Open primary sources such as SEC EDGAR and issuer disclosures form the baseline. Permitted market, calendar, transcript, and news sources extend the timeline. Collection may use direct APIs, bulk archives, scheduled fetchers, or Decodo when a public page needs reliable retrieval. A proxy changes transport reliability, not ownership or redistribution rights.
After acquisition, DataCedar retains raw evidence, hashes source objects, normalizes its own tables, checks expected coverage, and serves its own versioned API. Upstream adapters can change without becoming a runtime dependency for customers.
Deliberate boundaries
DataCedar is not a broker, exchange, investment adviser, or consolidated real-time SIP feed. Alpaca can validate historical rows during development, but it is not the public serving layer and is not used for present data. Text is served only where ownership or license permits it.
Questions about a source, use case, or coverage boundary? Contact DataCedar.