SWIRL for Backstage: Search That Gets the Results Right - Catalog, TechDocs, GitHub, Confluence

SWIRL for Backstage replaces the engine behind the Backstage search plugin. Backstage keeps its collators, its search page, and its permission filtering. SWIRL supplies a Tantivy index for the catalog and TechDocs, real BM25 relevance, and live results from GitHub and Confluence in the same ranked list. This demo runs on a fresh Backstage app with the published package and image. It starts with Backstage issue 27339, where the default engine matches "tech" to team entities, then shows the same query on SWIRL for Backstage, the hyphenated and partial-name cases, the three-step install, a GitHub result inside the catalog results, the scoped source in the SWIRL admin, and a container restart with the index still live on disk. 0:00 The default engine: "tech" returns team entities (issue 27339) 0:20 SWIRL for Backstage: one result, tech-radar 0:37 Hyphenated and partial names: foo-bar.com and abacus 1:00 Install: three steps and one container 1:28 GitHub results in the same list as the catalog 1:52 A scoped source in the SWIRL admin 2:16 Restart: the index lives on disk 2:40 Free, Apache 2.0, self-hosted Docs: https://docs.swirlaiconnect.com/backstage-overview Source: https://github.com/swirlai/swirl-backstage npm: https://www.npmjs.com/package/@swirl-search/backstage-plugin-search-backend-module-swirl Image: docker pull swirlai/swirl-backstage:0.1.1 Data stays in your systems, behind your permissions. Nothing is copied to SWIRL except what Backstage already collates.