onairos/docs/RAW_MEMORIES_SDK_UPDATE.md

The Onairos Library is a collection of functions that enable Applications to connect and communicate data with Onairos Identities via User Authorization. Integration for developers is seamless, simple and effective for all applications. LLM SDK capabiliti

# Raw memories — SDK update (handoff-aligned)

**Status:** Web SDK **8.3.15+**  
**Security:** Backend `authorizedData.rawMemories` + route guards (not optional policy).

---

## Backend contract (confirm with API team)

### `POST /getAPIurlMobile`

| Field | Type | SDK usage |
|-------|------|-----------|
| `authorizedData.rawMemories` | `boolean` | **Authoritative** — if `false`, SDK must not send `includeLlmData: true`, must not call `/llm-data/history`, and must not treat user UI approval as granted. |
| `authorizedData.llmData` | `boolean` (legacy) | Read only when `rawMemories` is absent (transition). |
| `rawMemoriesPolicy` | `{ offered: boolean, code?: string, connectorPlatforms?: string[] }` | UX: hide consent + raw connectors when `offered === false`. |
| `traitIntent` | `object` | Passed through on `onComplete` — SDK does **not** call `/api/trait-intents/resolve`. |

### Optional UI only — `GET /api/raw-memories/policy`

| Field | Type | SDK usage |
|-------|------|-----------|
| `rawMemoriesAvailable` | `boolean` | Prefetch before handoff to hide raw UI early (not security). |
| `code` | `string` | e.g. `APP_NOT_APPROVED_FOR_RAW_MEMORIES`, `RAW_MEMORIES_ALLOWED` |

Headers: `x-api-key` (required on handoff in production); optional `x-onairos-app-id` on policy prefetch.

**Trusted app id:** Backend allowlist uses `req.appId` from the developer API key, not client `Info.appId` / `webpageName`.

---

## SDK behavior summary

1. **Before handoff:** optional policy prefetch → `rawMemoriesOffered` state (`true` / `false` / `null`).
2. **Data Request:** hide Raw Memories row when `rawMemoriesOffered === false`; require explicit `requestData` / `rawMemoriesOnly` / `rawMemoriesConfig` when unknown (`null`).
3. **Onboarding:** strip raw-export connectors only when the host requested **raw export** (`rawMemoriesOnly`, `rawMemoriesConfig`, or `'rawMemories'` in `requestData`) **and** `rawMemoriesOffered === false`. **`personality` / `preferences` + `allowedPlatforms` including ChatGPT** keeps ChatGPT connect for server-side training; raw export stays blocked on handoff.
4. **Handoff request:** omit `LLMData` confirmation when `rawMemoriesOffered === false`.
5. **After handoff:** `includeLlmData` and history fetch gated on `authorizedData.rawMemories === true`.
6. **Exports:** `normalizeAuthorizedData`, `isRawMemoriesAuthorized`, `fetchRawMemoriesPolicyForUiOptional`, etc. from `onairos`.

---

## Integrator checklist

- [ ] Register app API key; do not rely on `webpageName` for raw access.
- [ ] Do not set `requestData={['rawMemories']}` unless Onairos approved the app (e.g. Juno).
- [ ] Read `onComplete` → `authorizedData.rawMemories` before using `apiResponse.llmData`.
- [ ] Use `traitIntent` from handoff / `onComplete`, not a separate resolve call.

---

## Parity

Mirror in React Native / Capacitor: same field names and gating order as `docs/CROSS_SDK_PARITY.md` § raw tier.