Introduction
Open-source TypeScript SDKs for official statistics, built for validated and reproducible data access.
Varve agency SDKs are open-source TypeScript clients for official statistical APIs.
They exist because public data APIs are often authoritative but difficult to use safely. Response shapes drift from documentation, identifiers are provider-specific, pagination rules differ, and error payloads are rarely consistent. These packages wrap those APIs in small, typed clients that validate every response before your application treats it as usable data.
The broader Varve thesis is simple: official numbers should be cited, reproducible, and refused when they cannot be proven. These SDKs are the provider-level access layer behind that idea.
What you get
- Typed clients for official statistical providers
- Runtime validation with Zod for every parsed response
- Isomorphic fetch support for Node.js 18+, browsers, Workers, and edge runtimes
- Provider-specific retries for transient agency failures and rate limits
- Exported schemas and types for downstream validation pipelines
- Documented quirks where live API behavior differs from provider docs
Packages
@varve/statcan-wds
Statistics Canada Web Data Service. Use it for tables, vectors, time series data, full table downloads, and release monitoring.
@varve/statcan-rdaas
Statistics Canada Reference Data as a Service. Use it for classifications, concordances, code hierarchies, and index lookup.
@varve/ons-api
UK Office for National Statistics. Use it for CDID time series, datasets, topic navigation, observations, and filtered exports.
@varve/eurostat-api
Eurostat APIs. Use it for JSON-stat data, SDMX metadata, SDMX CSV, dataset discovery, and graph metadata.
@varve/worldbank-api
World Bank Indicators API. Use it for countries, indicators, sources, topics, and country-period observations.
@varve/cmhc-api
CMHC Housing Market Information Portal. Use it for housing table exports, filter discovery, and normalized observation rows.
@varve/boc-valet
Bank of Canada Valet API. Use it for financial series, exchange rates, groups, observations, and FX RSS.
@varve/fred-api
Federal Reserve Economic Data. Use it for series, observations, releases, categories, tags, and GeoFRED data.
How the docs are organized
Start with Getting Started if you want to install a package and make a first request.
Use Choosing a Provider when you know the data problem but not the package.
Read the Core Concepts pages when you want to understand validation, retries, runtime behavior, and graph metadata across packages.
Provider sections contain the practical reference: data model, common workflows, configuration, errors, official references, and deeper guides where a package needs them.
Trust boundary
The SDKs verify the shape and transport behavior of official API responses. They do not decide whether a particular indicator, geography, seasonal adjustment, frequency, or transformation is analytically correct for your use case.
In other words: they make official data access more reproducible and easier to trust, while keeping provider ambiguity visible.