Table of Contents

Folder intake

Keep using your existing scanner or capture application by having it save completed documents to a folder that Paperweight watches. An enabled folder source processes supported files directly in that folder, without searching subfolders. Supported filename extensions are .pdf, .tif, .tiff, .jpg, .jpeg, and .png; Paperweight also checks the file's contents to confirm its format.

Choose a completion contract

Mode Capture application behavior Setup guidance
AtomicHandoff Write to a temporary filename, close it, then rename into the watched directory with its final supported extension Verify the producer never writes directly to an eligible final filename
ReadyMarker Finish and close the document, then create <document-name>.ready containing that document's 64-character SHA-256 hex digest Verify marker timing and hashes; a stale marker cannot authorize changed bytes
StableFile Leave size and modification time unchanged for stableSeconds, with write-excluding access possible Test slow saves and network delays: a pause in writing does not prove that a file is complete

StableFile defaults to ten seconds. For a new producer integration, prefer a demonstrable completion contract. A temporary extension must not be one of the eligible input extensions.

Ready-marker example

After the producer has closed this synthetic file:

$document = 'C:\PaperweightLab\Input\sample.pdf'
$digest = (Get-FileHash -LiteralPath $document -Algorithm SHA256).Hash
Set-Content -LiteralPath ($document + '.ready') -Value $digest -NoNewline

Do not use an empty sidecar. The marker records content identity, not only permission to scan a pathname.

Polling and source health

Paperweight checks folders every two seconds and can also respond to file-change notifications. Processing a document can delay the next check. If a folder is unavailable, Paperweight gradually increases the time between attempts, up to sixty seconds.

The Configuration tab shows source health. Ready means the source was successfully enumerated, not that every downstream document was accepted. Other states can identify unavailable sources, reconciliation needs, access failures, or a file-specific rejection.

Originals and duplicate notifications

Paperweight takes a protected local snapshot and records file identity and content hash before processing. Repeated notifications are not separate submissions of the same acquired file. New or replaced content is treated separately.

After publication, source cleanup verifies the original file's identity and content. A replacement at the same pathname is preserved. The local job directory then serves as the recovery archive, subject to retention.

Important

Assign each source to one workstation and Windows account. Do not enable competing consumers or point multiple sources at overlapping directories. Choose a completion mode that matches how your capture application finishes saving files.