Coverage

Ultralayer ingests the financial information flow across many channel types simultaneously, structures it into one coherent layer, and makes it queryable through every capability. This page describes the shape of that coverage.

Channels

Market-moving information doesn't arrive through one pipe. A story can break in a filing, get confirmed by a newswire, surface on an earnings call, and get dissected on a podcast hours later. We ingest across channel types so the platform sees the flow the way the market does:

Whatever the channel, everything lands in the same structured layer: entity grounding, scoring, classification, novelty clustering, and the timestamp discipline that makes it all backtestable. A filing and a newswire print are equally first-class citizens of the corpus.

Channel mix is curated deliberately and expands continuously — the criterion is signal, not volume.

Market focus

There is no hard coverage list — no fixed universe of tickers where coverage abruptly ends. Coverage is continuous, and it is shaped by two forces:

• A deliberate focus on US equities, particularly large caps. This is where our ingestion and curation attention is concentrated, and where coverage is densest.
• The natural breadth of the flow itself. Capabilities like the Wire ingest broadly across the global tape — geopolitics, commodities, currencies, crypto, rates, and international equities all flow through, because that is what the sources themselves cover. Wire items carry sector, industry, country, and asset-class tags precisely because the flow is global.

The practical takeaway: for US large-cap names, expect dense multi-channel coverage. Away from that center — small caps, international listings, niche markets — coverage exists and is often useful, but density varies with how much the global flow talks about it.

Historical depth

How far back the data goes varies from capability to capability — each product's corpus has its own history, and newer capabilities naturally have shorter ones. Rather than publish a table that would be stale in a month, the honest statement is: historical depth differs per product, and extending it is an active, ongoing commitment. We are continuously backfilling history — and because backfilled data receives simulated insertion delays rather than backdated timestamps (see point-in-time integrity), history added later is still safe to backtest against.

If your use case depends on a specific historical range for a specific capability, run a bounded query with start_timestamp/end_timestamp and see what comes back — the same point-in-time semantics that make backtests honest also make "what do you have for window W?" a directly answerable question.

Coverage is a moving target — upward

Coverage — channels, sources, breadth, and history — expands continuously. The best realtime picture of what's flowing is the platform itself: an unfiltered list_wire call shows you exactly what the wire is carrying right now.