Skip to content

“The index is a week behind the files.”

A search index that falls behind your files

Duplicate versions, re-embedding jobs and settings changes that need a full rebuild. DocSlurp treats every change as an update to what search already serves.

Why it happens

  • A re-upload lands as a new document, so the old version and the new one both answer.
  • Changing the embedding model or chunking means a rebuild, and search is down or wrong until it finishes.
  • Nobody can say which documents are indexed and which are still processing.

What changes with DocSlurp

  • Re-uploads become versions

    Upload the same filename to the same folder and it becomes a new version. Search and chat see only the newest; older versions stay readable until retention removes them.

    Details
  • Re-indexing that keeps serving

    A change that would make the index inconsistent queues the smallest re-index on its own. Until a document finishes, search serves it from its previous configuration.

  • Priced before it runs

    Every settings change is classified and estimated, so you see the documents, pages and cost a re-index will touch before you start it.

  • Know what is ready

    Stream each run’s progress, or get a signed webhook when it completes or fails.

    Details

sync.ts

const { run } = await client.upload({ files: [file], path: 'contracts/2026' });

for await (const snapshot of client.streamProgress(run.id)) {
  console.log(snapshot.run.status, snapshot.progress?.pageRollup?.completed);
}

Sound like your setup?

Tell us what you run today and what it costs. We’ll say plainly whether DocSlurp fits.

Jump to