Enterprise data warehouse

The data warehouse for hard-to-build datasets.

Rillor collects fragmented source data, turns it into structured and traceable records, and maintains it for enterprise analytics, AI systems, monitoring, and focused applications.

  • Collect
  • Label
  • Normalize
  • Maintain

Dataset registry

Data built around the evidence.

Each Rillor dataset declares what it contains, where it comes from, how it is structured, what it can support, and what remains limited. Access is confirmed by request, against current scope, source rights, evidence, freshness, and permitted delivery.

Legal Authority Corpus

Selected Brazilian administrative authority materials, source metadata, authority identity, document structure, and relationships

Source-linked research, authority retrieval, citation analysis, and legal-data infrastructure

GPU Systems, Pricing & Market Data

GPU system specifications, normalized configurations, listings, quotes, indicative prices, reported sales, verified transactions when available, and index-methodology research

Procurement comparison, market research, valuation and depreciation analysis, and governed index development

Federal Procurement Intelligence

Selected procurement, award, supplier, item, connector, and retrieval records

Supplier analysis, identifier resolution, sourcing research, and procurement-market mapping

Public Communications & Market Response

Selected communications, entities, classifications, timestamps, and time-aligned market observations

Event datasets, communications research, historical comparison, and methodology evaluation

Compare the full registry

Four current programs. Rillor’s scope is not limited to them.

The hard part

Source data rarely arrives ready for enterprise use.

Sources change. Records disagree. Identities drift. Formats break. Rights and provenance disappear. Rillor turns that unstable material into a dataset a downstream system can inspect and use.

Fragmented sources

The required records sit across many publishers, portals, filings, and formats with no shared index.

Inconsistent records and identities

The same entity appears under different names, identifiers, units, and field conventions.

Changing formats and content

Structures are revised and material is withdrawn, so one-time extracts quietly go stale.

Rights and provenance constraints

Origin context and source rights are lost during collection, leaving records unusable downstream.

Incomplete or conflicting observations

Observations disagree, and nothing records which reading was kept or why.

Custom datasets

Need a dataset that does not exist yet?

Define the sources, entities, fields, history, geography, and refresh cadence you need. Rillor can scope, build, and maintain the dataset around that requirement.

What you specify

Source universe

Which publishers, portals, filings, feeds, or first-party systems count as in scope.

Entities

The record and entity types that must be resolvable and joinable.

Fields

Required attributes, relationships, and the labels your systems expect.

History

How far back the record set must reach, where sources permit it.

Geography

Jurisdictions and markets the dataset has to cover.

Cadence

How current the data must stay, and what a change should trigger.

Delivery

How the finished records should reach the system that will consume them.

Custom work is subject to source feasibility, rights, scope, timeline, and commercial agreement.

How Rillor builds data

Collect. Label. Normalize. Maintain.

Every stage preserves the evidence the next stage needs: source context, identity decisions, transformations, quality review, versions, and known limitations.

  1. Collect

    Acquire and preserve source material with origin metadata and applicable rights information.

    • retained source artifact
    • capture context
  2. Label

    Identify records, entities, attributes, relationships, and evidence classes.

    • entity assignment
    • evidence class
  3. Normalize

    Reconcile structures, identities, units, terminology, and time without erasing the source record.

    • conformed schema
    • transformation lineage
  4. Maintain

    Version datasets, track changes, review quality, reconcile conflicts, and publish limitations.

    • dataset version
    • change and conflict record

Governance and verification apply across every stage — they are not a fifth step.

The full warehouse method

Managed delivery

Built for the system that will use it.

Delivery may be scoped as an authenticated API, scheduled export, or another agreed structured format. Schema, cadence, versions, rights, and support are defined with the engagement.

API access is a delivery method, not a standalone product. There is no anonymous or self-service public endpoint.

Authenticated API

Scoped endpoints with defined coverage, schema, authentication, rate limits, and versions.

Scheduled export

Recurring delivery on an agreed cadence, with version and change notes.

Structured file

An agreed structured format matching the receiving system's expectations.

Other managed arrangement

Another delivery arrangement suitable for your environment and controls.

Dataset standard

Every dataset should explain itself.

Rillor records provenance, rights posture, identity and schema decisions, lineage, evidence classes, freshness, quality, versions, and limitations. Every dataset record is written against that standard, so what a dataset holds and how it was built can be read off the record itself.

Source provenance

Where each record came from and under what capture context.

Source rights posture

The recorded rights position for the source material.

Record and entity identity

How records and entities are identified and resolved.

Schema and labels

The declared structure, field definitions, and label vocabulary.

Transformation lineage

What was changed between source record and dataset record.

Evidence classification

What kind of observation a value actually is.

Freshness and versioning

Declared refresh objective, dataset versions, and change history.

Quality and reconciliation

Review steps and how conflicting observations are handled.

Explicit limitations

What the dataset does not cover and should not be used for.

Tell us what data you need, how it will be used, and how current it must stay.

Request data

No account required · Text-only inquiry · Reviewed by the dataset team