NewWe now read the actual MRI and CT imaging — not just the radiologist's 1-page report. See how →
HomeProductNon-Clinical Image Filtering
Non-Clinical Image Filtering

Logos and blank scans never make it into the file.

Non-Clinical Image Filtering auto-detects images that carry no clinical content — letterhead logos, blank scan backs, cover art — and excludes them automatically before the file ever reaches review. The record set stays clinical, the page count reflects only pages that matter, and reviewers stop scrolling past cover art to find the next lab result.

Adams, T. · Case #IME-4812 342 pages
Images
Image
Letterhead logo, p.1 Excluded
Blank scan back, p.2 Excluded
Facility cover art, p.3 Excluded
X-Ray, right knee, p.88 Kept
MRI, right knee, p.91 Kept
Built on medical-record intelligence
Trained across legal, insurance & IME records, telling the page that documents care from the page that just came along for the ride.

Every image classified before it lands on the timeline.

As the file is ingested, each image is checked for clinical content — an X-ray, a photograph tied to a note, a slice from a DICOM study — before it's ever placed on the chronology. Letterhead graphics, fax banners, and blank page backs don't carry that signal, so they're routed out of the working set instead of onto the timeline.

Runs automatically during intake, on every upload
Page count reflects clinical content, not paperwork
Classification passOn upload
No clinical signal detected → excluded
Clinical imaging detected → kept, placed on timeline
Ambiguous or low quality → kept, flagged for review
Flagged for review2 items
Scan, low legibility, p.204 Review
Photograph, unlabeled, p.267 Review
Neither page is excluded automatically. Both stay in the working set until a reviewer confirms them.

Ambiguous never means dropped.

Filtering only ever removes what has no clinical content. A low-quality scan or an unlabeled photograph doesn't get excluded on a guess — it stays in the working set and gets flagged, so a reviewer decides instead of the filter.

Low-confidence images flagged, never silently dropped
Only non-clinical images are excluded automatically
Working page countBilled on Self-Service
Uploaded pages, 2 packets 342
Duplicates removed 11
Non-clinical images excluded Logos, cover art, blanks
The differentiator

A page count that means something.

On the Self-Service plan, pages are billed per page with duplicates already free. Logos and cover art don't get treated as records that need billing, review, or a spot on the timeline — they're filtered out before any of that happens.

See the Self-Service plan

From raw upload to a clinical-only file.

4 steps, no manual sorting through logos and cover art.

Upload the file

Any format, any volume. No pre-sorting of pages required.

Every image classified

Each image is checked for clinical content during intake.

Non-clinical images excluded

Logos, blank scans, and cover art are removed automatically.

Clinical images kept & flagged if needed

Ambiguous or low-quality scans stay in, flagged for your review.

Who keeps their file clinical-only.

Whoever sends the raw packet, the reviewer gets the record without the wrapper.

Related features.

FAQ

Non-clinical image filtering, answered.

Letterhead logos, blank scan backs, cover art, fax banners, and other pages that carry no clinical content. They're detected automatically and excluded from the working file so they don't sit inside the chronology alongside actual imaging and notes.

No. Filtering targets non-clinical images only. Clinical images, including low-quality or ambiguous scans, are kept in the working set and flagged for review rather than silently dropped.

Each image is classified for clinical content before it's placed on the chronology. A letterhead graphic, a blank page back, or cover art doesn't carry the signal of an X-ray, MRI slice, or photograph tied to a clinical note, so it's routed out of the working set instead of onto the timeline.

On the Self-Service plan, pages are billed per page with duplicates already free. Excluding non-clinical images from the working set keeps the page count centered on content that matters to the case, rather than letterhead and blank scans.

Yes. It runs as part of intake, the same pass that segments and OCR-routes every page, so filtering happens before the chronology is built rather than as a manual cleanup step afterward.

Yes, they run on different content. Non-Clinical Image Filtering targets images without clinical value. Cover Page & Certification Stripping targets law-firm cover sheets and custodian certification pages. Both keep the working set focused on the clinical record.

They're removed from the working view that feeds the chronology and reports, not deleted from the source upload. The original file is preserved.

See your file with the logos already gone.

Book a quick demo, or upload a single file and see what gets filtered before you review a page. No commitment. Handled under our BAA; never used to train a model.