Cryptographic verification is now live. Govern, attest and mint every check.Go Now

1:N Biometric Matching

One face.
Every database.

Q: What is 1:N biometric search?

1:N search compares a single face against an entire database of enrolled faces to find matches. Unlike 1:1 matching (which compares two specific faces), 1:N search identifies who someone is from a large population — used for duplicate detection, watchlist screening, and fraud ring identification.

Q: How accurate is deepidv biometric matching?

deepidv achieves a 99.7% true acceptance rate with less than 0.01% false acceptance rate. Performance is consistent across diverse demographics, lighting conditions, and appearance variations including aging, glasses, and facial hair.

Q: How fast is 1:N database search?

Sub-second search times for databases up to 1 million enrolled faces. Results are returned as a ranked list of matches with confidence scores, allowing your system to process and act on results in real time.

Q: Does biometric matching include liveness detection?

Yes. All biometric matching includes active and passive liveness detection to prevent spoofing with photos, videos, masks, or deepfakes. Liveness is verified before any matching comparison begins.

Q: How is biometric data stored and protected?

Biometric templates are mathematical representations of facial features — not images. Templates cannot be reverse-engineered into photos. All biometric data is encrypted at rest and in transit, with full compliance for GDPR, CCPA, BIPA, and biometric privacy regulations.

Advanced facial recognition for 1:1 identity verification against ID photos and 1:N search across databases of millions of enrolled faces — with liveness detection, fraud-ring detection, anti-spoofing, and configurable confidence thresholds.

Book a demo

DocumentationDeveloper docs

deepidv verifies one face against an ID, searches it across millions of enrolled faces, and flags duplicates before they ever enroll.

1:1 match

99.7similarity

1:1 identity verification

Match a live selfie to the ID photo with a 99.7% true-acceptance rate — handling aging, glasses, and facial hair across diverse demographics.

1:N search2.4M

1:N search at scale

Search databases of millions of enrolled faces in under a second and get back a ranked list of matches with confidence scores.

DedupDuplicate

≈

Similarity0.98

Catch duplicates & fraud rings

Surface the same face across multiple accounts or identities — stopping duplicate enrollments and connected fraud before they spread.

By the Numbers

Built for real-world accuracy

Enterprise-grade biometric matching with sub-second 1:N search and advanced liveness detection.

99.7%
True acceptance rate: <0.01%
False acceptance rate: <1s
1:N search response: ISO 30107-3
Presentation attack detection

1:1 Verification

Confirm a face is who it claims

Capture a live selfie, extract a biometric template, and compare it to the reference photo from an ID or your database. A calibrated confidence score returns a pass/fail you can branch on — tuned to your own risk tolerance.

Where 1:1 fits

Customer Onboarding

Selfie-to-ID at account creation

Transaction Auth

Verify high-value actions

Account Recovery

No knowledge-based questions

Age Verification

Confirm age against the ID

Employee Verification

Secure facility & remote access

Return to Service

Re-verify after risk signals

deepidv · 1:1 verifymatching

compare

Live selfie

ID reference

99.7sim

Identity confirmed

Templates match · liveness passed

Above threshold · 0.90

Feature points compared68 landmarks · 128-d template

1:N Search

Find one face across millions

Submit a probe face, generate its template, and search every enrolled face in your database. Sub-second performance returns all matches above threshold, ranked by confidence — so you can identify who someone is, not just confirm a claim.

What 1:N unlocks

Duplicate Detection

Catch synthetic-identity fraud

Watchlist Screening

Known fraudsters & banned users

Fraud Ring Detection

One face, many accounts

Access Control

Identify without credentials

Event Security

Screen against VIP & watchlists

Lost Person Search

Locate or identify individuals

deepidv · 1:N searchscanning 2.4M

Probe template128-d · extracted0.6s · sub-second

Ranked candidates

tmpl_8841a2

96.1

tmpl_2290ff

54.3

tmpl_5512cc

48.7

tmpl_3397bb

41.2

1 match above thresholdthreshold · 0.90

Liveness & Anti-Spoofing

Match only living, unique people

Active and passive liveness runs before any comparison begins, and ISO 30107-3 presentation-attack detection blocks photos, replays, masks, and deepfakes. Automatic deduplication then flags the same face enrolling twice under different identities.

Protection layers

Active & Passive Liveness

Confirm a real person

Presentation Attack Detection

Photos, replays, masks

Deepfake Detection

Synthetic-media defense

3D Depth Analysis

Screen-presentation defense

Face Quality Assessment

Blur, lighting, pose, occlusion

Encrypted Templates

Math, never reversible to a photo

Read the docs

deepidv · dedup checkduplicate

Duplicate enrollmentBlocked

Same biometric template already enrolled under a different identity

acct_8841tmpl_8841 · Submitted now

Jun 22

acct_3190tmpl_3190 · Enrolled earlier

Mar 04

Template similarity0.98

Trusted & Certified

Built to meet the world's most demanding security and privacy standards.

SOC 2 Type II

Audited security controls for data protection

GDPR Compliant

Full compliance with EU data privacy regulations

ISO 27001

International standard for information security

DIACC Member

Digital ID & Authentication Council of Canada

Integration

Take full control.
Getting started takes minutes.

Plug deepidv into your stack three ways — a REST API, native SDKs, or an MCP server your AI agents can call directly. Clean docs, a sandbox, and one source of truth take you from API key to live verification the same day.

REST API

Call any verification service directly — create a session or run a standalone check and get structured JSON back in under 500ms.

→ screening 200+ watchlists…
✓ status: clear
  risk_score: 4 · 143ms

Drop-in SDK

Pre-built UI components and server libraries for iOS, Android, Web, Node, Python, and Go — zero to live verification in a single sprint.

  checks: ['id','liveness','face'],
✓ v.url → send to your user

MCP for agents

An open protocol that lets any AI agent call deepidv's verification engine directly — no custom integration code. Connect it in one command.

Claude

Launch deepidv for my AI agent — set it up from the docs.

✓ 12 verification tools available

Ready to run ID, liveness & AML checks.

Changelog

The Latest

See all releases

May 18, 2026Release

Verify 2.0 is here

Multi-language verification, a redesigned ID & face-capture flow, end-to-end whitelabel, and a native Background Check step.

May 9, 2026Features

ProofCall, smoother capture & translated reports

AI background-confirmation calls, faster camera start-up, translated PDF/HTML reports, and location-by-IP enrichment.

April 10, 2026AI

AI-powered adverse media screening

Continuous, AI-driven adverse-media checks built natively into your verification workflows.

See all releases

Frequently asked questions

1:N search compares a single face against an entire database of enrolled faces to find matches. Unlike 1:1 matching (which compares two specific faces), 1:N search identifies who someone is from a large population — used for duplicate detection, watchlist screening, and fraud ring identification.

deepidv achieves a 99.7% true acceptance rate with less than 0.01% false acceptance rate. Performance is consistent across diverse demographics, lighting conditions, and appearance variations including aging, glasses, and facial hair.

Sub-second search times for databases up to 1 million enrolled faces. Results are returned as a ranked list of matches with confidence scores, allowing your system to process and act on results in real time.

Yes. All biometric matching includes active and passive liveness detection to prevent spoofing with photos, videos, masks, or deepfakes. Liveness is verified before any matching comparison begins.

Biometric templates are mathematical representations of facial features — not images. Templates cannot be reverse-engineered into photos. All biometric data is encrypted at rest and in transit, with full compliance for GDPR, CCPA, BIPA, and biometric privacy regulations.

Still have questions?

Our team is ready to help you get started.

Contact Sales

All-In-One Products

Explore the Full Suite

Chain Verification

Face Liveness

1:N Biometric Matching

Address Verification

AML Questionnaires

One face. Every database.

See deepidv biometric matching on your own data — it takes 60 seconds to start.

Get Started View API Docs

One face.
Every database.