Skip to main content

AI Lens

Search 10 million files. Get the right one. In 3 seconds.

AI Lens runs on every backup. Find a face across 10 years of photos. Search text in scanned PDFs. Identify objects in photos. View DICOM medical images. CCTV recognition. Document digitization.

Six Modes. One Backup.

All AI Lens features at a glance.

Faces & People

Train on a single photo, find them across all backups. 512D embedding, pgvector-powered.

Image Labels & Objects

EfficientNet-Lite0 labels every photo. Search "beach sunset" or "dog wearing sunglasses."

OCR & Text Search

PaddleOCR runs on every PDF, screenshot, scanned doc. Full-text search.

Medical (DICOM)

DICOM viewer + metadata extraction (patient ID, modality, study date). Cornerstone3D-powered.

Read more

CCTV & Surveillance

Recognize people, vehicles, time-of-day patterns in surveillance footage. (Coming soon.)

Read more

Digitization

Scanned documents become searchable databases. Bank statements, receipts, contracts. (Coming soon.)

Read more

How It Works

AI runs on the backup. Not on your endpoint.

When a file lands in BigMind, our AI Worker pipeline picks it up — face detection, embedding, labels, OCR — and writes results to the search index. No CPU drain on your machines. Models run on our infrastructure.

Models used

SCRFD (face detect) + ArcFace (embedding) · EfficientNet-Lite0 (labels) · PaddleOCR (text) · Cornerstone3D (DICOM). All ONNX-runtime-based; ~410MB of models.

Where It Shines

Three example use cases.

Photographer

David Mojica: finds client photos by name across 10+ years.

Healthcare

DICOM search by patient ID + modality across years of imaging studies.

Insurance

Find every photo from a claim filing date even if filenames are useless.

Pricing

Where AI Lens lives in pricing.

TierFaces / Labels / OCRDICOM MedicalCCTVDigitization
DR-Only
Standardsoonsoon
Prosoonsoon

CCTV and Digitization are in active development — coming soon.

AI runs on every backup. So search just works.