Knowledge Unified Katalog And Network
A Data Catalog for Everyone
Read, query, connect. A data platform for people and AI.
Publishing is where it starts. KUKAN stays with the data until it is actually used.
Know what a file holds without opening it, and search right inside it.
Register a file and an AI reads it and writes a short description. You also get title, description and tag suggestions to check and adopt field by field.
Learn moreHybrid search, by keyword and by meaning, finds the data even when it uses different words. You see which file matched and why, across CSV tables and PDF text alike.
Learn moreView CSV, PDF, Office, Ichitaro, Markdown and more right in the browser. Tables sort and filter in place, and GeoJSON shows on a map.
Learn moreAsk the data a question before you download and open it.
Query a tabular resource in SQL from the browser or over the API. Pull just the rows and columns you need and aggregate on the spot, JSON lists included.
Learn moreReplacing a file keeps the old content as a version. For a table, see the rows added, removed and changed, and delete a version that leaked personal data.
Learn moreExternal-URL resources are checked on a schedule, and broken links show up in the dashboard. Changed content is picked up again, so previews and search keep up.
Learn moreNot only through the web UI: from AI agents, BI tools and the tools you already have.
A built-in MCP server lets Claude and other AI agents find data in natural language. They cite documents down to the chapter, section and page, and query tables in SQL.
Learn moreA tabular resource carries an OData connection URL. Paste it into Excel, Power BI or Tableau to open the table, aggregate and chart it, with nothing to download.
Learn moreA CKAN-compatible read API answers package_search and the rest as-is. Migrate while your CKAN clients and links with other catalogs keep working.
Learn moreContainers deploy to AWS with the bundled CDK and pipeline, or on-premises with Docker Compose. Air-gapped networks such as LGWAN work too.
Learn moreSeveral catalog sites share the hourly-billed boxes: the database cluster, OpenSearch and the load balancer. Inside them each site keeps its own database, search index and bucket, so an added site costs little.
Learn moreGive members roles per organization and category (group) to decide who edits each dataset. A private dataset is visible only to members of its organization, and API tokens carry the same permissions.
Learn moreAGPL-3.0 with the full source open, so anyone can deploy it and adapt it. The UI and docs switch between Japanese and English, and colors, logo and wording can be made your own.
Learn moreComing up: SSO (OIDC) sign-in, settled column types for CSV carried into diffs, and overriding column roles.
If your organization runs KUKAN, let us know in a GitHub issue. It helps us decide what to build next, and we may ask to list your logo and a link on this site.
Tell Us You Use It
