Coverage and depth

Coverage and historical depth

Dilution wedge — now live in all 4 SDKs and CLI (April 2026). Eleven row types (events, warrants, convertibles, ROFR, lockups, cash position, corporate actions, Nasdaq compliance, ratings, reverse splits, share-float history) plus a coverage rollup and a per-issuer score, exposed via 14 REST endpoints (/v1/dilution/*), 14 MCP tools, and first-class methods in the TypeScript, Python, Go, and Rust SDKs (and omni-sec dilution * subcommands). Thirteen of the 14 routes honor ?view=agent; coverage is already compact. See Dilution endpoints.

This page summarizes the current public coverage posture without overstating what Datastream has not yet shipped. The right mental model is:

• a broad public workflow surface built around SEC and adjacent market utilities
• a 9-year searchable history floor for filings and section workflows
• stronger completed-year completeness guarantees for the most recent two full years of filings and section workflows
• a 30-year cold-storage policy target that remains intentionally deferred from the hot path
• explicit partial and gated semantics where a surface is not fully generalized yet

XBRL facts coverage

• 95M+ pre-parsed XBRL facts stored in PostgreSQL across 9,600+ SEC-registered companies
• all SEC XBRL taxonomies covered: us-gaap, ifrs-full, dei
• daily automated sync via Dagster pipeline at 07:00 UTC refreshes the full universe
• DB-first query architecture: facts and statement endpoints query our database first, falling back to SEC EDGAR only when data is missing

Current snapshot

• filing and section history validations currently pass across the 9-year searchable floor
• the release gate separately enforces stronger per-form completeness floors for the last two completed years instead of pretending all 9 searchable years carry the same density guarantee
• historical coverage is proven through checked-in validation artifacts, not a generic “full archive is always hot” claim
• current v3 public workflow coverage includes:
  • entities and canonical resolution
  • filings, latest filing retrieval, and section extraction across 10-K / 10-Q / 8-K / DEF 14A plus offering forms (S-1, S-3, 424B*, FWP, 20-F)
  • normalized statements, facts, segmented revenues (with geographic depth detection and hierarchy filtering), share float, and common equity derivation with preferred stock deduction
  • ownership, insiders, historical institutional extract, and compensation
  • SEC-native M&A events, SEC-furnished earnings materials, enforcement events, and board composition
  • market calendar, rights-scoped indices, and OMNI Volatility Score
  • market data: snapshots, historical bars, corporate actions, reference, estimates, and search
  • factor intelligence: 84-factor catalog across market, style, macro, sector, and industry categories with daily returns, correlations, intraday, dashboard, and regime performance
  • macro intelligence: indicators, releases, calendar, forecasts, high-signal packs (12 countries including 24 daily US FRED series), and regime classification
  • investment intelligence: security, company, earnings preview, portfolio, watchlist, semantic query, and footnote analysis
  • portfolio strategies: factor decomposition, stress testing, optimization, factor rotation, and regime screening

Measurement coverage

• the freshness dashboard tracks current measured forms for launch-critical SEC workflows
• the history inventory tracks the searchable filings plus sections floor separately from freshness
• the benchmark suite tracks repeated investor-agent workloads separately from freshness so speed claims do not blur into freshness claims

Partial or intentionally scoped surfaces

• market calendar: XNYS and XNAS are holiday-configured; other covered exchanges remain weekend_only
• index constituents: NDX is public; SPX, DJIA, and RUT remain rights-scoped for public export even though internal mapping coverage is strong
• offerings: public filing records are live; section-level extraction (Prospectus Summary, Risk Factors, Use of Proceeds, Capitalization, Dilution, Description of Securities, Underwriting, Selling Stockholders, Plan of Distribution, Principal Stockholders, Management, Executive Compensation, Related-Party Transactions, 424B pricing supplements, FWP term sheets, 20-F items) is now live for S-1 / S-3 / 424B* / FWP / 20-F with a dedicated catalog and exhibit parsing for EX-4, EX-5, and EX-10 (warrant, legality, and material-contract exhibits). Richer LLM-extracted dilution payload (warrant terms, lockup schedules, rights of first refusal) remains deferred to the dilution wedge workstreams that build on top of this extractor.
• sentiment: intentionally gated; not public
• traceability: filing-derived document-plus-page support is live, HTML-first exact region anchors now ship for supported segmented revenue facts and supported section excerpts, and the shared trace surface now also covers SEC enforcement source records; broader PDF/cell parity and richer non-filing families remain deferred
• analytics: a narrow tenant-safe query surface is live across filings, sections, segmented revenues, SEC enforcement actions, and quarter-specific institutional ownership extracts; raw SQL and broader warehouse-style coverage remain deferred

Historical depth posture

• searchable hot posture today: 9-year filings and sections floor
• cold posture target: 30-year archive
• non-goal for v3: full 30-year hot normalization

The repo still keeps warm and cold tiering as an explicit backlog item rather than pretending that policy targets are already the same thing as current production implementation.

Source artifacts

• ops/history-inventory/latest.json
• ops/history-quality/contract.json
• ops/history-quality/latest.json
• ops/external-reference-validation/latest.json
• ops/freshness-dashboard/latest.json
• docs/strategy/omni-datastream-v3-evidence-matrix.md