Live Pipelines, Faster Reports, Better Sign-in

This release focuses on clearer operational workflows and a smoother day-to-day experience. You will see faster reporting in key areas, plus new pipeline views that make it easier to understand what is ready and what needs attention.

Smoother performance in reporting and discovery

Several backend-heavy screens and reports now do less repeated work while keeping results exact. This helps dashboards and large datasets feel more responsive without changing what you see.

• Reduced database work for processing history, directory counts, topic retention, and inference reports, while preserving exact totals.
• Workload reporting now focuses on active jobs and jobs completed within the selected reporting window.
• Broad recommendation queries and source filtering do less repeated visibility and filtering work.
• Source tag identity lookups are reused until the tag catalog changes.
• Completed analysis catalog indexes are released instead of being retained in memory.

New operational workflows and reliability fixes

Admins can now manage source onboarding and article readiness using live pipeline boards. These views support targeted retries and readiness checks, helping prevent activation or publication before required processing is complete.

Partner imports are more flexible, with tier-based sync intervals and support for larger catalogs through configurable page limits. Import and assessment failures are also easier to act on, with clearer diagnostics and more accurate product status labels.

Sign-in and sign-out behavior is more consistent across the reader, admin, partner portal, and Chrome and Edge extensions. Sign-out now ends the provider SSO session, provider logout is honored across apps, and failed or canceled sign-ins can be retried while preserving the intended destination.

Temporary API outages no longer have to strand users on broken pages. Source, topic, and tag pages now remain recoverable during upstream failures, with navigation and retry available on web and extensions.

This release also expands public discovery for automated clients, including reader API documentation, sign-in provider metadata, OAuth resource metadata, and MCP server capabilities through standard discovery URLs.

Finally, a few smaller quality fixes are included, such as consistent DevFeed icons in the partner portal, clearer article attribution that distinguishes original headlines from DevFeed previews, and publishing article tags as search keywords.