Align Oral-History Transcript Timecodes Web Application
Correct transcript timings using supplied recording anchors.
Category Directory
Check and prepare local AI models, digital collections and digitized materials.
Correct transcript timings using supplied recording anchors.
Update recognized text while retaining page relationships.
Create structural descriptions of digitized objects.
Organize successive collection versions into structured preservation objects.
Sample multiple sources into a documented dataset mixture.
Package queries, documents, and supplied relevance judgments.
Inventory model metadata and structural file problems.
Inventory archived installer packages without running them.
Find missing resources within supplied website captures.
Measure signal defects across recorded audio batches.
Find malformed dialogue and tool-call structures.
Locate structural problems in archived database file sets.
Identify structural adapter/base mismatches.
Verify that expected access copies exist and link correctly.
Find missing shards, invalid offsets, and inconsistent tensor mappings.
Find missing or misplaced pages in digitized volumes.
Find scans outside declared technical specifications.
Inspect supported captures for selected signal-pattern problems.
Create traceable text chunks using saved splitting rules.
Check selected content properties across migration batches.
Compare metadata and tensor layouts across model versions.
Evaluate explicit decoding candidates for legacy collections.
Work with locally selected EML or MBOX exports, including supported messages inside ZIP files. This app processes the files in your browser and does not connect to a live mailbox. Compare two exports by sender, recipients, dates, body content, and attachment hashes; distinguish meaningful losses from formatting or MIME-encoding differences.
Compare file collections within supported optical-disc images.
Find stale, missing, or orphaned indexed content.
Identify tensor additions, removals, and structural changes.
Locate tokenizer changes affecting encoded text.
Transform conversation records without silently discarding fields.
Transfer text and supported layout structures between formats.
Apply repeatable negative inversion and color settings.
Build or check standards-based transfer packages.
Build presentation metadata for ordered collection objects.
Turn supplied preservation events into structured metadata.
Create frame files linked to their original sheets.
Preserve supported compiled help content as browsable files.
Recover allocated files from supported disk-image layouts.
Detect reserved names, case-only duplicates, trailing dots, Unicode collisions, and long paths. Export a portable ZIP plus a reversible rename map.
Work with locally selected EML or MBOX exports, including supported messages inside ZIP files. This app processes the files in your browser and does not connect to a live mailbox. Use supplied project/customer references and message relationships to propose groups where people changed the subject; export reviewed conversation packages.
Compare two local mail exports and build a small MBOX containing only missing messages, with duplicate and conflicting copies separated.
Select archived responses using capture metadata.
Locate frame boundaries and produce numbered image collections.
Identify dependencies inside archived Office documents.
Calculate model-specific token lengths for local datasets.
Calculate target-patch differences under explicit conditions.
Repackage web archives with rebuilt indexes and collection metadata.
Build fixed-length token blocks with explicit masks.
Gather model files and supporting configuration into reproducible releases.
Preserve evaluation inputs, predictions, settings, and results together.
Keep OCR regions synchronized with edited page images.
Match scanned files to expected physical collection items.
Deduplicate records while retaining their origins.
Reorganize tensors into new files without changing their values.
Keep related records together when creating dataset partitions.
Check multimodal dataset references and usable image content.