> ## Documentation Index
> Fetch the complete documentation index at: https://docs.truu.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Data handling

> TruU acts as your data processor: personal data from a verification belongs in storage your organization owns, and TruU keeps only the facts of what happened. What TruU reads, stores, masks, and deliberately does not store.

## TruU is your data processor

Your organization is the controller of your employees' personal data. TruU IDV processes it on your behalf, only to run the verification you asked for, and is built not to hold on to it. **Personal content from a verification belongs in storage your organization owns, not in TruU.**

| | Where it lives |
| - | - |
| **Personal content:** fields read from an ID document, the verified selfie used as a face reference, document images and selfie video | Your own Amazon S3 bucket or Azure Blob container, written with your credentials and, if you choose, encrypted with your key |
| **Facts about the verification:** which factors ran, how each concluded, when, and who requested it, keyed to the directory identifier with contact values masked | TruU, so the request and its audit trail can be reconstructed. Audit events exclude PII |
| **Nothing at all:** camera frames, one-time codes, knowledge answers, vendor session media, directory copies, passwords | Not stored by TruU anywhere |

Connect your store in **Settings → Storage** before you turn on the Proofing factor, and no document field, selfie, or face reference is ever written to TruU. See [Bring your own storage](#bring-your-own-storage).

> [!WARNING]
>
> Until you connect your own store, proofing records and face references are written to TruU managed storage, partitioned for your organization and held on your behalf. Connecting a store later and choosing **Move TruU data** moves those objects into your store and out of TruU.

## Principles

TruU IDV is built on a data-minimization principle: **store enough to reconstruct what happened; do not store enough to recreate the verification artifact.** TruU records that a verification occurred, which factors ran, and how each concluded, without becoming a second copy of your directory or a repository of biometric material.

## What TruU IDV reads and stores

### Directory data

TruU IDV connects to Microsoft Entra ID with read-only access through a customer-owned app registration. See [Microsoft Entra ID](https://docs.truu.ai/idv/connect/directory/entra-id.md).

* Directory attributes are read to drive factor availability, knowledge questions, and verification delivery.
* Attribute metadata (which fields are populated, and coverage percentages) is cached to power schema discovery; TruU does not replicate your full directory.
* Group memberships are looked up live at verification time; only the list of group names and IDs for the use case rule picker is cached (five minutes, refreshed daily).
* User passwords and tenant administrator credentials are never accessed or stored.

### Verification outcomes

For each verification request, TruU IDV stores the structured record an auditor or incident responder needs:

* Who requested it (for example, the help desk agent), for whom, and when.
* Which use case and flow applied.
* Each factor attempted, its outcome, and a reason code on failure.
* Completion state (verified, failed, expired, cancelled) and timing, and which directory attribute served each factor.

### Audit records

Every meaningful admin, agent, and verification action lands in a tenant-scoped, append-only audit trail. The audit model, including its redaction rules, is documented in [Audit events](https://docs.truu.ai/idv/security/audit-events.md).

## What is stored from proofing and face

Each of these is a setting on the Proofing factor, so you decide what a document check leaves behind. Everything in this section that contains personal data is written to your own storage once you connect a store.

* **Proofing decision record.** On by default. When a document-and-selfie check completes, TruU writes a decision record containing the outcome, the check results, the name match result, and the identity fields read from the document. It is written to your own storage when you have connected a store; without one, it falls back to TruU managed storage partitioned for your organization.
* **Capture media.** Document images and selfie video from the proofing session are archived only if you turn on capture archiving in the proofing factor's settings. It is off by default.
* **Verified selfie as a face reference.** Off by default. When **Proofing photo** is on, the selfie from a passed proofing check is deposited as the user's face reference so later verifications can use the face factor without a directory photo.
* **Vendor-side data.** For the integrated provider, TruU deletes the vendor session after the decision record is written. For Persona, PingOne Verify, and Entrust, retention is governed by your contract with that vendor.
* **Face verification.** Live face checks produce a match score and liveness confidence; the comparison runs in AWS Rekognition and the captured frames are not retained by TruU after the result is recorded. Reference photos live where you configure them: your directory, your storage, or TruU-managed storage.

## What is not stored

* **One-time codes.** Submitted OTP values are never written to verification or audit records.
* **Directory copies.** No mirror of your users or memberships; attribute coverage statistics and a group catalog only.
* **Passwords and tenant admin credentials.** Never accessed.
* **Face images by default.** Proofing face references and capture media are opt-in only. See above.
* **Directory write access.** The connection is read-only; there is nothing for TruU to modify in your tenant.

## PII masking and redaction

Sensitive values that do appear in records are masked by default:

| Value | Stored / displayed as |
| - | - |
| Phone number `+1 555 123 4567` | `(***) ***-4567` |
| Email `user@domain.com` | `u**r@d*****.com` |
| Document image | Not stored unless capture archiving is enabled |
| Face match score | Confidence band in consoles and audit; the numeric score is kept on the verification record for review |
| OTP code | Never stored |

The end-user contact preview returns masked values only: phones show the last four digits, and emails keep the first and last character of the local part plus the domain initial. In the help desk app, agents sending a link see the email and mobile on file (and can edit them) so they can confirm the delivery target with the caller; audit records store contact values masked.

Audit records are held in TruU's database, not in object storage, and are not affected by the bring-your-own-storage setting; export them to your SIEM to keep a copy you control. Audit events themselves exclude PII. The stored event carries identifiers and masked display fields only; the full event payload, where it contains sensitive detail, is held in controlled object storage that the event references, under the same tenant isolation and encryption as other verification assets. Every audit event is flagged to indicate whether PII redaction was applied, and visibility into sensitive fields is governed by role. See [Audit events](https://docs.truu.ai/idv/security/audit-events.md).

## Storage and secrets

### Bring your own storage

This is how personal data stays out of TruU. An owner or admin connects your organization's own cloud store, and every verification asset that carries personal data (decision records, face references, capture media) is written there instead of to TruU:

* **Supported backends:** Amazon S3 and Azure Blob Storage.
* **How to connect:** **Settings → Storage → Add store** in the IDV Console. See [Storage](https://docs.truu.ai/idv/connect/storage.md).
* **Once connected,** new assets are written to your bucket using your credentials, never TruU's. You own the bucket, its access policy, its encryption, and its lifecycle.
* **Moving existing data out of TruU:** choose **Move TruU data** on a Write store to make it your primary store and move anything already in TruU managed storage into it. Assets stay readable throughout the move.
* **Without a connected store,** these assets fall back to TruU managed storage, partitioned per organization.
* Every storage change is recorded in the audit trail.

You can define several named stores with Read and Write access (for example, read face references from an existing HR photo bucket while writing proofing records to a separate bucket), and optionally encrypt each object with your own key before it is written. Objects are named so your own lifecycle rules can manage them.

### How credentials and secrets are protected

* **Storage credentials are write-only.** You can set them but no API ever returns them.
* Stored secrets (storage, directory, SSO, connector) are held in a managed secrets service.
* **Superseded secrets are deleted** from the secrets service when you rotate or reconfigure.
* API client secrets and keys are stored hashed and shown once, at creation or rotation.
* Token-signing private keys are never persisted; only public keys are published.

## Encryption

* **In transit:** all traffic between the verification page, the consoles, and the TruU IDV API runs over TLS.
* **At rest:** stored verification assets and audit records are encrypted at rest.

For details of TruU's broader encryption and infrastructure controls, see [Compliance](https://docs.truu.ai/idv/security/compliance.md) or contact TruU for current security documentation.

## Retention

* Retention of verification and audit records is set by your agreement with TruU.
* For long-term retention, export the audit trail (CSV or JSON Lines) into your own SIEM or data lake. Once exported, retention is governed by your own rules.
* Where document and selfie proofing runs on the integrated provider, the vendor session is deleted after completion, and the decision record lives in storage you control or in TruU-managed storage under your retention terms. Proofing run through Persona, PingOne Verify, or Entrust is governed by your own contract with that vendor. See [Proofing](https://docs.truu.ai/idv/factors/proofing/overview.md).

For the controls that support BIPA and GDPR programs, see [Biometric privacy controls](https://docs.truu.ai/idv/security/compliance.md#biometric-privacy-controls).

## Tenant isolation

TruU IDV data is tenant-scoped throughout. Every event, verification record, and configuration object carries a tenant identifier, and every query (console search, API access, and export) is scoped to the requesting organization. Asset references are validated against the owning organization on every read; a reference pointing at another tenant's data is treated as not found. Help desk agents are further scoped below the tenant level: they see audit data only for the verification requests they are handling, not the tenant-wide trail.

## Your controls

* **Keep personal data in your own storage** by connecting your own S3 bucket or Azure Blob container, so TruU holds only the facts of each verification.
* **Revoke directory access at any time** by removing or deleting the app registration in your tenant.
* **Govern who sees what** with console roles: owners and admins manage configuration; agents operate request-scoped.
* **Own your evidence** by exporting audit events into systems you control.
* **Disable factors or attributes** you do not want used for verification (for example, excluding a poorly maintained attribute from knowledge questions).

## How data moves

A verification request moves through at most six parties. Not every request touches every party: a flow built from SMS, email, and knowledge factors never reaches the proofing vendor or AWS Rekognition.

```mermaid theme={null}
flowchart LR
    U["User device<br/>(browser on phone or desktop)"]
    T["TruU IDV<br/>orchestration, transient"]
    V["Proofing vendor<br/>document check"]
    R["AWS Rekognition<br/>liveness and face comparison"]
    S["Your storage or<br/>TruU managed storage"]
    X["SIEM or ticket<br/>(ServiceNow, export)"]

    U -- "link, codes, answers" --> T
    U -- "document and selfie capture<br/>(vendor SDK)" --> V
    U -- "streamed face video" --> R
    T -- "session, flow, use case" --> U
    T -- "start check, poll outcome" --> V
    V -- "outcome, extracted fields, selfie" --> T
    T -- "reference photo, compare request" --> R
    R -- "liveness result, match score" --> T
    T -- "decision record, face reference" --> S
    T -- "audit events, request result" --> X
```

What happens at each hop:

1. **User device to TruU.** The user opens the verification link. TruU resolves the request, applies the use case, and serves the flow. One-time codes and knowledge answers are submitted to TruU and checked there; submitted codes are never written to any record.
2. **User device to proofing vendor.** For the Proofing factor, the vendor's SDK runs in the user's browser and sends document images and the selfie directly to the vendor. TruU does not proxy the capture. The vendor returns the outcome, the identity fields it extracted from the document, and the selfie image to TruU.
3. **User device to AWS Rekognition.** For the Face factor, the user's camera video is streamed to AWS Rekognition for liveness. TruU then asks Rekognition to compare the live frame against the reference photo. Rekognition returns a liveness result and a match score; nothing from the session persists in AWS once the result is returned.
4. **TruU to storage.** When a proofing check completes, TruU writes the decision record to your storage if you have connected one, otherwise to TruU managed storage. When **Proofing photo** is on (on in new workspaces), the verified selfie is deposited as the user's face reference in the same place. Capture media is written only if **Store capture videos/images** is on (integrated provider only, off by default).
5. **TruU to SIEM or ticket.** Every step lands in the audit trail. The ServiceNow connector pushes lifecycle events to your instance; the audit export API carries the trail to your SIEM.

TruU itself is transient in this flow: it holds session state while the request is open and keeps the audit record and the request outcome after it closes. It does not keep camera frames, submitted codes, or vendor session media.

## Custody after the request closes

| Party | Holds | Does not hold | Retention | Governed by |
| - | - | - | - | - |
| **TruU** | Audit events and the verification request record (who, for whom, which flow, each factor outcome). With TruU managed storage: the proofing decision record and the deposited face reference, partitioned per organization. Capture media only when capture storage is on. | Submitted one-time codes, camera frames from face checks, vendor session media, directory mirrors, passwords | Set by your agreement with TruU. Export audit events to keep your own long-term copy. | Your TruU agreement and the [Data handling](https://docs.truu.ai/idv/security/data-handling.md) commitments |
| **AWS Rekognition** | Nothing after the result is returned | Streamed video, face images, match scores, reference photos | None | TruU's agreement with AWS; AWS is a TruU subprocessor |
| **Proofing vendor (integrated provider, Didit)** | The vendor session for the duration of the check. TruU requests deletion as soon as the decision record is written | | Governed by TruU's agreement with the vendor | TruU's agreement with Didit; a TruU subprocessor |
| **Proofing vendor (your own: Persona, PingOne Verify, Entrust)** | Whatever your vendor retains under your contract: document images, selfie, extracted fields, vendor transaction records | Nothing TruU adds; TruU sends the session reference and reads the outcome | Set by your vendor contract | Your contract with the vendor; the vendor is your processor, not TruU's subprocessor |
| **Your storage** (Amazon S3 or Azure Blob) | Proofing decision records, deposited face references, capture media when capture storage is on, encrypted with your key if configured | Audit events (these stay in TruU and reach you through export) | Your bucket lifecycle rules | You |

> [!NOTE]
>
> Connecting your own storage changes where the decision record and face reference are written. It does not move the audit trail. Export audit events to keep a copy you control; see [Audit events](https://docs.truu.ai/idv/security/audit-events.md).

## Controls that change the flow

| Setting | Where | Effect on custody |
| - | - | - |
| Storage: Your stores | **Settings → Storage** | Decision records and face references are written to the store you choose on the Proofing factor instead of TruU managed storage |
| Decision record | Proofing factor settings | On by default. Off: no decision record is written anywhere; only the outcome stays on the request |
| Store capture videos/images | Proofing factor settings (integrated provider) | On: document images and selfie video are written alongside the decision record. Off by default |
| Proofing photo | Proofing factor settings | On in new workspaces. On: the selfie from a passed check is deposited as the user's face reference |
| Face reference source | Face factor settings | Chooses whether the reference photo is read from your directory, your storage, or TruU managed storage |

## Related pages

> [card] [Audit events](https://docs.truu.ai/idv/security/audit-events.md)
>
> The event model, redaction flags, and immutability guarantees.

> [card] [Compliance](https://docs.truu.ai/idv/security/compliance.md)
>
> How these controls support your compliance program.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.