Platform

Six modules. One audit trail. All composable.

Each Passflow product runs standalone or composes into the full pipeline. Same SDKs, same dashboards, same decisioning model.

Identity Verification Document Verification Liveness Detection 1:1 Face Match AML / Sanctions Developer SDKs
01 / Identity Verification

End-to-end KYC, decisioned in under two seconds.

Document capture, OCR, biometric match, AML screening and decisioning — orchestrated through one workflow. Drop into your onboarding flow with three lines of code.

The problem we solve Most enterprises in regulated industries have stitched together 4-6 vendors for KYC. Each one owns a piece of audit log. When the regulator asks "show me how this account passed," you spend two weeks pulling reports from four dashboards.
Verification · vf_8a13c2
CT · 12:04 PM
✓
Document captured
US Driver's License · CA · MRZ valid
0.4s
✓
Liveness check
PAD Level 2 · passive · score 0.987
0.3s
✓
1:1 face match
Selfie ↔ document · 99.82% similarity
0.2s
✓
AML / Sanctions screening
OFAC, PEP, adverse media · no hits
0.4s
→
Decisioning
Risk model us_retail_v3 · approving
0.1s
99.6%
Pass-through accuracy
1.4s
Median end-to-end
−87%
Synthetic ID fraud
POST /v1/verifications
# Start an identity verification
curl https://api.passflow.ai/v1/verifications \
  -H "Authorization: Bearer $PASSFLOW_API_KEY" \
  -d '{
    "workflow": "us_retail_kyc_v3",
    "applicant": { "reference": "acct_28104" },
    "residency": "us-ashburn-1"
  }'
02 / Document Verification

10,000+ document types across 200+ countries.

MRZ parsing, NFC chip read, security feature detection, tamper analysis. Updates ship monthly from a team of forensic document analysts.

The problem we solve Document libraries go stale within months as states reissue IDs. Most vendors update quarterly, miss new templates, and your fraud team eats the false rejects.
CALIFORNIA
DRIVER LICENSE
DL B6483102
CHEN, MARIA E.
DOB 03 14 1992
SEX F · EYES BRN
ISS 09 22 2024
Security feature analysis
MRZ checksum✓ PASS
UV overlay pattern✓ PASS
Hologram detected✓ PASS
Microprint integrity✓ PASS
Ghost image match✓ PASS
Edge tamper analysis0.04 risk
OCR ↔ MRZ name✓ MATCH
10k+
Document templates
200+
Countries supported
Monthly
Template update cadence
03 / Liveness Detection

iBeta PAD Level 2. One frame. No wiggle gestures.

Passive liveness from a single selfie. No "turn your head" choreography that loses 20% of your applicants. Active liveness available as fallback for high-risk segments.

The problem we solve Active liveness adds 30-90 seconds and a 15-25% drop-off. The new generation of deepfakes can pass three out of four legacy "blink and turn" tests.
PAD Level 2 · score 0.987 · GENUINE
CAPTURE · 1080×1920 · 30fps
● LIVE
98.7%
PAD level 2 score
1frame
Required input
<300ms
Inference time
04 / 1:1 Face Match

NIST FRVT top decile. Auditable similarity scores.

Selfie-to-document, selfie-to-ID-photo, selfie-to-existing-customer. Returns a similarity score with model version and decision threshold — what your audit team needs.

The problem we solve Most face-match vendors return a binary pass/fail. Examiners want to see the threshold, the model version, and the score distribution for the workflow.
SELFIE · LIVE CAPTURE
REFERENCE · DOC PORTRAIT
99.82%
SIMILARITY SCORE
Modelface-match-v4.2
Threshold0.94
DecisionMATCH
Demographicparity ✓
Top 10%
NIST FRVT ranking
0.03%
False-match rate at 0.94
Audited
Demographic parity report
05 / AML / Sanctions Screening

OFAC, PEP, adverse media. Continuous monitoring.

Screen on apply. Re-screen on schedule. Re-screen on event. Built-in case management for analyst review. 250+ lists curated and updated daily.

The problem we solve Most banks bolt on a watchlist vendor that fires alerts into a separate console. Cases get lost between the KYC team and the BSA team. We unify the timeline.
Case · cs_4a82c1 · Maria E. Chen
No hits
OFAC SDN
UN Consolidated
EU Sanctions
HMT (UK)
PEP · Global
Adverse media
OFAC SDN screening result last run · 12:04 PM
CANDIDATESCOREDOB DELTADISP
Chen, Maria0.71+11 yrCLEARED
Chen, Maria L.0.65−4 yrCLEARED
— no further candidates above 0.6 —
Next scheduled re-screen: 2026-06-13 Trigger on: address change · sanctions list update
250+
Watchlists, refreshed daily
<400ms
Screening latency p95
14events
Re-screen triggers built-in
06 / Developer SDKs

Five platforms. Sandbox in five minutes.

iOS, Android, Web, Flutter, React Native. REST API + signed webhooks. A test bank account with mock data so you can wire it up before the contract closes.

The problem we solve Most vendors gatekeep their SDK behind a sales call. Your CTO can't evaluate. We give read-only access to the sandbox the day you sign an MSA — even before the SOW.
iOS
Swift 5.9+
SPM · CocoaPods
Android
Kotlin · Java
minSdk 24+
Web
React · Vue
ES2020 · TS
Flutter
Dart 3+
pub.dev
React Native
0.72+
npm · yarn
POST /v1/verifications · receive webhook
import Passflow from "@passflow/node"

const pf = new Passflow(process.env.PASSFLOW_API_KEY)

const verification = await pf.verifications.create({
  workflow: "us_retail_kyc_v3",
  applicant: { reference: "acct_28104" }
})

// Receive a signed webhook when decisioned
app.post("/passflow", pf.verifyWebhook, (req, res) => {
  if (req.body.event === "verification.approved") {
    openAccount(req.body.applicant.reference)
  }
  res.sendStatus(200)
})
Ready to evaluate?

Run our full pipeline against your fraud sample.

Book a benchmark → Read the docs ↗