Every filing,
as structured data.
XBRL-driven intelligence on 2.7M UK private companies. API and MCP, ready for your pipeline.
2.7M+
API-ready companies
Up to 10 yrs
Financial history
100%
XBRL-sourced
The cleanest interface for XBRL-driven data.
Structured financials, ownership chains, and company indicators — extracted from Companies House filings, ready for your pipeline.
Three views of the same filing.
Structured financials. Ownership. Indicators. Extracted from XBRL, not scraped.
Financials
Ownership
Indicators
Cycle through 5-7 years of financials, ownership data, and key indicators.
Clean, parsed JSON responses extracted directly from raw XBRL filings.
Predictable, machine-readable schemas designed to feed LLMs, financial models, and compliance workflows.
One endpoint. Every team that touches UK company data.
Same API. Different query. Click a role to see what it looks like in practice.
Shortlist across 2.7M+ UK companies by revenue, ownership, and sector.
// 30 targets: UK industrials, £10–50m rev, founder-owned
const targets = await perpetua.companies.search({
sector: "28.*", // SIC: machinery
revenue: { min: 10e6, max: 50e6 },
ownership: { type: "individual", min_pct: 50 },
limit: 30,
});Priced like infrastructure, not a magazine subscription.
Start free. Scale by usage. Enterprise terms on request.
Developer
Up to 1,000 calls / month
Explore, prototype, evaluate.
Professional
Per seat / month · billed annually
Teams in production.
Enterprise
Bulk, on-prem, audit-grade
Regulated workloads, SLAs.
All tiers include TLS and the same underlying dataset. SOC 2 Type II audit in progress; reports available once the audit period closes. Pricing excludes VAT.
UK private company data should be structured, provenance-linked, and trusted by default.
The UK publishes more machine-readable financial data than almost any country in the world. Companies House receives nearly every limited company's accounts in XBRL — a structured, tagged format. And yet most platforms serve this data as scanned PDFs, scraped HTML, or hand-typed summaries.
We think that's a bug, not a market. Analysts spend days reconciling what the regulator already knows. Credit teams chase ownership chains across five tabs. AI agents hallucinate revenue figures that sit one API call away in plain text.
Perpetua is one endpoint over the full XBRL corpus. Every figure links to the filing it came from. Every ownership chain traces back to the PSC register. Every number is what the regulator filed, not what a model guessed.
We are built for deal, credit, compliance, and accountancy teams who need the answer to be right — and for the next generation of AI agents that will run the same queries at scale.