> For the complete documentation index, see [llms.txt](https://docs.openg2p.org/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.openg2p.org/platform/platform-services/vc-issuance/phase-2-device-wallet.md).

# Phase 2 — Device Wallet

Phase 2 (future) — self-owned smartphone wallets (Inji Mobile). The self-sovereign upgrade for citizens who have a smartphone.

Phase 2 adds **self-owned smartphone wallets** for the growing segment of citizens who have a smartphone. It is the **self-sovereign** complement to Phase 1's paper credential — the citizen holds their own keys and credentials on their device.

> This is a forward-looking section. Phase 1 (paper) is the implementation focus; Phase 2 is added as smartphone adoption in the target population grows.

## Carried over — capabilities deferred from Phase 1

These are **not** wallet features; they were consciously sequenced out of Phase 1 and are picked up here so nothing is lost track of.

| Deferred capability                | Why it was deferred                                                                                                                                           | What taking it up involves                                                                                                                                                                                                                                                                                                                                                                                                                                         |
| ---------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| **Photograph in the claim-169 QR** | Sequenced behind the core issuance chain.                                                                                                                     | Request the photo as part of the **eSignet KYC** response (the ID system's own photograph, matching what the beneficiary authenticated against), downscale it to a **\~1–2 KB** WEBP/AVIF thumbnail, and push it as the **`face`** claim so Certify embeds it in the signed QR. Needs a hard size budget: the whole QR is capped at **\~2.9 KB**, shared with the signature and any embedded certificate. Certify never fetches images — the bytes must be pushed. |
| **Revocation / status lists**      | A paper credential is verified **offline**, so a status list cannot be consulted at scan time; short credential validity is the Phase-1 compensating control. | Certify already ships the status-list tables (`status_list_credential`, `credential_status_transaction`, `status_list_available_indices`) and `allowed-status-purposes={'revocation'}`. A device wallet **can** be online at presentation, so revocation becomes genuinely checkable — which is why it belongs here. Decide the posture (revocation vs suspension) and who may revoke.                                                                             |
| **Android agent app**              | The reference **web** portal came first, against the same API.                                                                                                | A native client adds field-grade capability for roaming agents — most importantly **Bluetooth printing**, which a browser cannot do. The eSignet redirect is handled with a Custom Tab plus polling of the authentication status, so no deep-link plumbing is required.                                                                                                                                                                                            |

## What it adds over paper

* **Holder-bound presentation** — bound to a key on the device, so a credential can't simply be photocopied and reused (unlike bearer paper).
* **Online / remote presentation (OpenID4VP)** — present digitally to a remote relying party, not only face-to-face.
* **Offline presentation** — QR/BLE in the field, no connectivity needed.
* **Selective disclosure** — reveal only the claims a verifier needs.
* **Self-service** — fetch, hold, and re-present without a kiosk trip.

## Component

* **Inji Mobile** — MOSIP's OpenID4VCI device wallet (Android/iOS). Any OpenID4VCI-compliant wallet also interoperates, since Certify is standards-based.

## Same issuer, registry data pulled (not pushed)

Phase 2 reuses the **same Inji Certify issuer** — only the **holder/delivery** changes (a device wallet instead of paper). The data path, however, flips: instead of the Phase-1 **push** (where the Agent Portal API reads the Registry and pushes claims), the wallet does an interactive **OpenID4VCI** download, so **Certify pulls** the citizen's claims itself via the custom **Registry connector** (`RegistryDataProviderPlugin`), keyed by the citizen's authenticated token claim. See [Registry Data Connector](/platform/platform-services/vc-issuance/registry-data-connector.md).

That wallet flow needs an OAuth2/OIDC **authorization server** and introduces the integration considerations (token type, `c_nonce`, the AS choice) that Phase 1 deliberately avoids. Those are taken up when Phase 2 is scheduled.

## Not this: the agent-assisted handover

A credential can already be delivered into a citizen's wallet **at the counter**, by handing them an OpenID4VCI credential offer — see [Wallet Handover (Phase 1.5)](/platform/platform-services/vc-issuance/wallet-handover.md). That reuses Phase 1 entirely: same agent, same authentication, same **push**, same single Certify.

It is worth being clear that this is **not** Phase 2 and does not remove the need for it. It gives a smartphone owner holder binding and digital presentation, but the citizen still travels to an agent. Phase 2's defining property is **self-service** — the citizen downloads unaided — and that is what requires the pull connector and a second Certify configured against eSignet.

## Not in scope (separate tracks)

* **Hosted/custodial wallet (Option B, Inji Web + Mimoto)** — considered but **not chosen**; see [Custody Options & Strategy](/platform/platform-services/vc-issuance/custody-options-and-strategy.md). Adopt only as a deliberate custodial-locker policy.
* **Consent-based data sharing** (department ↔ department / third-party pull) — a registry partner-API + consent concern, not a wallet feature.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.openg2p.org/platform/platform-services/vc-issuance/phase-2-device-wallet.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
