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

# Memory and personalization

> Learn how Praxa remembers, adapts, explains learned context, and keeps correction or consent with the owner.

Learn how Praxa remembers, adapts, explains learned context, and keeps correction or consent with the owner.

<Note>
  Persistent memory and AHCE are Praxa product capabilities. Use @praxa/sdk/memory for read-only provider federation; AHCE is not a public ingestion endpoint.
</Note>

## Provided capabilities

| Capability                                                                                                                                                                      | Status | Surfaces                   | Who it helps |
| ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------ | -------------------------- | ------------ |
| **Persistent agent memory**<br />Retain durable facts, preferences, and context with user-facing correction and deletion controls.                                              | Live   | iOS, web, voice, on-device | Individual   |
| **Goal and task capture**<br />Propose explicit goal or task records from cloud chat and keep the owner in control of acceptance.                                               | Live   | iOS, web                   | Individual   |
| **Adaptive persona and tone**<br />Adapt tone and framing to the person without changing factual or authority boundaries.                                                       | Live   | iOS, web, voice            | Individual   |
| **Adaptive Human Context Engine**<br />Use consented, bounded context signals to improve situational relevance; this is a Praxa product capability, not a public ingestion API. | Live   | iOS                        | Individual   |
| **Adaptive Context transparency and correction**<br />Show learned rhythm information and let the person correct or abstain from it.                                            | Live   | iOS, web                   | Individual   |
| **Proactive suggestions**<br />Surface consent-gated suggestions that require an explicit tap before any next action.                                                           | Live   | iOS                        | Individual   |
| **Native onboarding**<br />Guide a new user through sign-in and initial setup in a compressed native flow.                                                                      | Live   | iOS                        | Individual   |
| **Persona recipes**<br />Apply a prepared persona preset for a recurring role such as meal or trip planning.                                                                    | Live   | iOS, web                   | Individual   |

## Example outcomes

1. **Correct a durable memory.** Open Memory settings, change or delete an owner-visible memory, and verify future recall respects the correction.
2. **Capture a goal.** State a concrete goal in cloud text chat and review the proposed goal record before accepting it.
3. **Review adaptive context.** Open Adaptive Context, inspect the learned rhythm, and submit a correction or abstention.
4. **Federate an existing store.** Use the memory federation tutorial to query an existing provider without migrating it.

## Verify the surface

1. Confirm the capability status and listed surface in the [Capability Atlas](/overview/capabilities).
2. Sign in to the intended Praxa product surface and select the correct personal or organization workspace.
3. Exercise the example with the smallest safe scope and inspect the visible result or approval card.
4. Test one unavailable, denied, or disconnected state so the surface fails honestly.
5. Use a developer tutorial only when this guide links to a supported public developer contract.

<Card title="Integrate existing memory" icon="database" href="/memory-federation/overview">
  Use read-only federation when your backend already owns a supported memory provider. This does not write into AHCE or Praxa personal memory.
</Card>