- eKYC Platform
- Face Authentication
- Vision Score
- Behavioral Biometric (BehavioIQ)
- SIM Verification
- Mule Account Database
- eKYC Core
- Demo Apps
Overview
TrustVision is our best-in-class AI-driven computer vision technology, with proprietary models for face matching, optical character recognition, and liveness check, training our models on emerging market data for industry-leading accuracy. Through these technologies we provide an eKYC (Digital KYC) solution to promote financial inclusion and access, serving 5 emerging markets and over 500M+ customers: Vietnam, India, Indonesia, the Philippines, and Bangladesh. Our team comprises 22+ PhDs and 23+ M.S. in AI and Data Science with 15+ years of experience in machine learning and AI, with research published at NIPS, ICML, ICLR, and KDD.
This site documents three ways to integrate with TrustVision, plus a set of independently callable products. If you're not sure where to start, read the comparison below, then see each product's own section in the sidebar (Overview, Integration Use Case, API Docs, and SDK Integration where applicable) for deeper guidance on choosing.
Choosing an integration path
| eKYC Core API | eKYC Platform | Standalone products | |
|---|---|---|---|
| What it is | The original, single-purpose API surface (upload an image, compare two faces, read an ID card, verify an NFC chip, etc.) that you compose yourself into a full identity-verification flow. | Bundles the full eKYC journey — document verification and selfie/liveness verification — into a single, pre-configured flow selected via flow_id. | Independently callable APIs for a single capability, used standalone or alongside either of the above. |
| Auth | HMAC request signing. | AES-256 encrypted_token (Platform API) / JWT (Service API). | Varies by product — see each product's docs. |
| Who it's for | Clients integrated before the eKYC Platform existed. Remains fully supported. | New integrations that want a faster path — the journey (which steps run, in what order, against which checks) is pre-configured per client. | Clients who only need one capability (e.g. just a trust score, or just face re-authentication) without a full eKYC flow. |
eKYC Core
The eKYC Core API is TrustVision's original, HMAC-authenticated API surface — a collection of single-purpose endpoints that clients compose themselves into a full identity-verification flow. It predates the eKYC Platform and remains fully supported. It's organized into Image/File, Face, ID Card, NFC, Transaction, and Metadata sub-sections, plus Authentication and the composite Customized API.
eKYC Platform
The eKYC Platform bundles the full eKYC process — document verification and selfie/liveness verification — into a single, ready-made journey, selected via flow_id and pre-configured for your account. Because the steps that run, and the order they run in, are already decided, integrating against it is typically faster than orchestrating each check yourself against the eKYC Core API. Integration follows three steps regardless of journey: start a session (Authentication), launch the journey via a mobile SDK, WebView, or a direct Platform API call (or the bundled Onboard call), then retrieve the result from the JWT-authenticated Service API. See the eKYC Platform overview for the full walkthrough.
Standalone products
These are independently callable — use one on its own, or alongside either API above:
- Vision Score — assesses a customer's trustworthiness and risk profile from a selfie image.
- Face Authentication — register a customer's face once, then re-authenticate against it on future sessions.
- Behavioral Biometric (BehavioIQ) — device registration, behavioral data collection, and fraud scoring based on device and interaction behavior.
- SIM Verification — confirms ownership of a phone number and surfaces telco-held SIM information.
- Mule Account Database — report and search for accounts suspected of money-mule activity.
For a deeper walkthrough of how to choose between these, see the Overview page in each product's own section (e.g. eKYC Platform).