Overview

The vertically integrated platform serious investors have been missing.

Most tools stop at data, stop at chat, or stop at dashboards. OMNI is built around the full path: source-backed intelligence, an agent that can use it properly, and a workspace that makes the result usable every day.

Explore OMNI Datastream Explore OMNI Agent Explore OMNI Pulse

Grounded by public-source data and filing lineage

Backed by scoped dated benchmark proof, not universal black-box claims

Built for investor-agent workflows rather than generic chat

Designed around the market day investors actually run

How it works

Three layers, in sequence.

The stack gets better because each layer improves the next one instead of making it compensate for missing structure.

Step 1

Datastream turns public-market sprawl into a dependable substrate.

Entities, filings, sections, statements, ownership, governance, fund holdings, calendars, traceability, and derived utilities come out structured, compact, and source-backed.

Step 2

Agent turns substrate into repeated work.

Prompt routing, skills, tools, evals, and memory help with research, monitoring, and synthesis instead of just generating a paragraph.

Step 3

Pulse turns repeated work into a better day.

Markets, portfolio context, briefings, terminal views, media, and embedded chat stay in one workspace that already knows the context.

Features

The platform is strongest when the layers reinforce each other.

One layer by itself is useful. All three together are much harder to replace.

The foundational public-market data layer

Datastream is the substrate. If the substrate is messy, everything above it stays brittle.

Canonical entity graph

Resolve issuers, filers, managers, insiders, funds, and aliases across ticker, CIK, FIGI, ISIN, and CUSIP without rebuilding identity logic in every client.

Filing search that fits agents

Search by ticker, form, accession, date range, and sort order with compact metadata and predictable pagination.

Section and item extraction

Pull the filing slices that matter instead of dragging an entire document through every workflow.

Statements and facts

Use normalized statements, facts, bundles, segmented revenues, and share-float reads with provenance attached.

Ownership and insider reads

Track 13F, insider, beneficial ownership, and ticker-holder workflows from a single data layer.

Fund and governance reads

Pull N-PORT holdings, board composition, and enforcement-event surfaces from the same public contract family.

Historical depth

Keep hot-path reads fast while holding a deeper historical corpus behind the same object model.

Freshness metadata

Every serious downstream system needs to know how fresh the answer is. OMNI makes that explicit.

Parser and provenance metadata

Materialization version, source URLs, accession references, request IDs, and trace references stay close to the payload.

Filing-page traceability

Go from a filing-derived number to the correct filing document and page instead of hand-hunting through PDFs.

Rights-safe market utilities

Market calendars, approved index rosters, and derived public-market utilities live alongside SEC data.

Hosted MCP

The same primitives are exposed to coding agents without forcing a custom wrapper first.

API, SDK, and CLI parity

The platform is built so the docs, reference, SDKs, and programmatic surfaces reinforce each other.

Performance

Performance you can see. Not just a line in a deck.

OMNI wins where the repeated investor-agent workflow is the real benchmark: resolving entities, searching filings, returning compact facts, and preserving the metadata that downstream agents need.

latency