01 — PipelineData system

Build a source-backed company record

Discover sources, resolve entities, and timestamp every claim before analysis begins.

01
Trace a claim
Open the source, excerpt, confidence, and date.
02
Resolve identities
Join founders, companies, patents, and aliases.
03
Reuse clean data
Skip extraction when a document has not changed.

01 / pipeline

retrieval → resolution → graph

stages

companyfoundersprior companiestechnical claimscompetitorscustomerscapitalIPresearchhiringreleasesregulatory exposurepartnershipsgeographymarket structure

retrieval

Forty sources per company pass, each one dated and held against the claim it produced.

cohortsfilingspatentsreposjob posts

03 / data graph

fig. 1 — every edge carries ⟨src · date · conf · span · contradiction⟩

typed edges11

provenance fields5

contradiction states3

02 / entity resolution

01 John K. Smith @ Startup X

02 J. Smith — former DeepMind

03 J.K. Smith, third author, NeurIPS 2021

04 Smith, John Kenneth — inventor, US 11,842,013

→ one canonical node · person#4471

01.2 — Routing

Structured model routing

Keep raw webpages out of final reasoning.

Fast models structure sources. Specialist reasoning receives only data, contradictions, and open questions.

Luna · synthesis receives

verified factsfounder graphproduct graphmarket graphcompetitor matrixIP analysistechnological trajectorycustomer signalscapital historyhiring signalscontradictionsunusual observationstimelineopen questions

The ratio is the point. Up to forty source extractions collapse into one analysis and one act of judgment. Luna receives structured data rather than raw webpages, which keeps the expensive context focused and the output auditable.