About OMB.report

OMB.report is an unofficial public reference site and is not operated by OMB or any U.S. government agency.

Use OMB.report as a public reference and discovery layer. For official records, authoritative language, and agency actions, consult the originating federal source.

The site links OMB control numbers, ICR reference numbers, IC collections, and related public documents under stable public URLs.

Recent data is synced hourly. Historical backfill, verification, and repair run continuously in the background until the archive converges.

Primary metadata is collected from Reginfo.gov. Popular-download trend data comes from public federal analytics feeds. Binary documents are stored in managed object storage and streamed through this domain so headers and availability can be controlled consistently.

Wrapper pages remain the preferred citation target because they preserve filing context, identifiers, related links, and availability state.

Public document wrappers live under this domain. When binary artifacts are available, OMB.report streams them with app-controlled headers rather than redirecting users to object storage.

When a document or preview is missing, the site surfaces the metadata honestly and queues repair work in the importer rather than pretending a preview exists.

Recent and historical backfill may lag the upstream source during large repair or reindexing windows. Some historical records can remain partial until background verification completes.

Search, sitemap, and document coverage are continuously rebuilt from the importer state, so coverage improves over time instead of staying silently stale.

When the site shows partial metadata, missing previews, or an incomplete document bundle, the importer queues a repair target and retries the missing artifact in the background.

For production operations, use the importer health report and sync logs first; request-time page repairs are a safety net, not the primary update path.

Use the government directory, agency list, popular downloads, and feeds to browse the corpus without relying on search alone.