Cleaned Swedish open data as an API

A simpler way to handle swedish open data api. Early access is coming.

Product spec: Cleaned Swedish open data as an API

Who it is for

Developers and small product teams building apps, dashboards or internal tools on Swedish public data.
Journalists and researchers who need reliable data without writing scrapers or cleanup scripts.
Consultancies and civic tech groups that repeatedly combine data from several Swedish agencies.
Teams building AI tools that need structured, trustworthy Swedish public data.

The problem

Sweden publishes a lot of open data, but each agency does it differently. The Riksdag, Skolverket, SMHI, SGU and Sveriges Radio all have their own formats, identifiers, date conventions, pagination and documentation, often in Swedish only. Municipality and region codes are not always consistent across sources. Endpoints change or break without warning.

So anyone who wants to combine sources, for example school results by municipality alongside local weather or parliamentary decisions, has to write and maintain a separate cleanup layer for each source. Most teams rebuild this same work from scratch.

Five must-have features

1. One consistent API across sources. Every source uses the same conventions: JSON over HTTP, the same pagination, filtering, sorting and error format. Learn it once and use it for every dataset.

2. Cleaned and normalized data. Standard ISO dates and times, consistent units, correct encoding of Swedish characters, deduplicated records, and official municipality and county codes wherever geography appears. This lets users join datasets without extra work.

3. Provenance and freshness on every record. Each response shows the original source, the original endpoint, when it was fetched and what cleaning was applied. Users can trust the data and cite it.

4. Clear schemas and documentation in English and Swedish. Field descriptions, example queries, and a machine-readable OpenAPI spec so client libraries can be generated for any language.

5. Bulk export alongside the query API. Full dataset downloads in CSV and Parquet for analysts, plus a changelog of schema changes so integrations do not break silently.

What to leave out (for now)

Too many sources at launch. Start with a small set that is often combined, such as Riksdag, Skolverket and SMHI, and do them well before expanding.
Dashboards, charts and a visual front end. This is infrastructure, not an analytics product.
A chat interface or built-in AI features. An MCP layer can come later as a thin wrapper over the API.
Real-time streaming. Scheduled refreshes with visible timestamps are enough for most use cases.
Personal or restricted data. Only data that is already officially published as open data.
Language-specific SDKs beyond one reference client. The OpenAPI spec covers the rest.

How it is different from the competitors

The competitors found are all free, open-source projects on GitHub. They differ from this product in four ways.

Single source vs. many. Each competitor covers one source or one domain. Riksdag-Regering-MCP and RiksdagenPythonAPI cover parliament and government data. Skolverket-MCP covers education data. sondera covers SMHI and SGU hydrology and meteorology. sr-traffic covers Sveriges Radio traffic messages. None of them normalize data across agencies so it can be joined. That cross-source cleaning is the core of this product.

Tied to one environment vs. language-agnostic. RiksdagenPythonAPI and sondera are Python packages. The two isakskogstad projects are MCP servers built for LLMs. sr-traffic is an Angular app, not a data service. This product is a plain HTTP API that works from any language, tool or AI agent.

Wrappers vs. maintained, cleaned data. As far as their descriptions show, the competitors mostly pass through or parse data from the original APIs. This product adds normalization, provenance, freshness tracking and schema change notices.

Hobby projects vs. a maintained service. The star counts (5 to 31) suggest small, niche projects. This is not a criticism of their quality, but users who build products on top of public data need stable uptime and someone responsible for fixes. Reliability is part of the offer.

Note: zsith/launcher.user.js is a userscript for the game agar.io and has nothing to do with Swedish open data. It looks like a false match and should be dropped from the competitor list.

Watch-out: The same author is behind both Riksdag-Regering-MCP and Skolverket-MCP, which suggests a deliberate push toward broader coverage of Swedish public data for LLMs. That is the closest thing to a direct competitor. It may also be a potential partner, for example as an MCP front end on top of this API.