Skip to content
EN
English 简体中文 soon 日本語 soon

TurboLens

Visual forensics APIs for image fraud and extraction

Visit official site

What TurboLens is

An API layer for document and image trust. Placed after upload and before approval workflows, it answers two distinct questions: has this existing image been locally edited, and was this image generated or modified by AI?

Forgery detection returns a verdict, the proportion of suspicious pixels, region coordinates and a heatmap. Downstream, specialized OCR extracts structured data from invoices, identity documents and contracts, with particular depth in Southeast Asian formats.

What you can do with it

  • Screen images for splicing, copy-move edits and retouching
  • Check whether images are AI-generated or AI-edited, in early access
  • Extract structured data from invoices, receipts and passports
  • Route flagged submissions to manual review by your own thresholds
  • Integrate via REST JSON without rebuilding your pipeline

Who it is for

  • KYC, lending and insurance teams verifying submissions
  • Document processing pipelines needing authenticity checks
  • Platforms fighting fraudulent identity uploads
  • Engineering teams adding forensics without in-house CV work

What to watch out for

  • Synthetic-media detection is early access; treat results as screening, not proof
  • Automated verdicts support review but do not replace human judgment or business rules
  • Handling identity documents makes compliance and retention terms critical
  • Accuracy varies with image quality and manipulation sophistication

Pros & cons

✓ What we like

  • Two distinct forensic questions kept separate and traceable
  • Evidence outputs (heatmap, regions) support human review
  • Regional OCR depth for SEA documents
  • Straightforward API integration

! What to watch out for

  • Detection is probabilistic and can be evaded
  • Early-access features carry change risk
  • Identity data handling demands strict compliance

FAQ

What is the difference between forgery and AI detection here?

Forgery detection finds edits to an existing image; AI detection screens whether generative AI created or modified it. TurboLens treats them as complementary layers.

Where does it fit in a workflow?

Typically immediately after upload and before OCR, extraction or approval, with flagged items routed to manual review.

What evidence does it return?

A real-or-forged verdict, tampered-pixel proportion, suspicious-region coordinates and a heatmap.

Last reviewed: 2026-09-15

More AI office assistant tools

View all →

How we review