> For the complete documentation index, see [llms.txt](https://tinyhumans.gitbook.io/openhuman/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://tinyhumans.gitbook.io/openhuman/features/settings.md).

# Settings

What each Settings panel controls, where the autonomy tier lives, and which old addresses now redirect into Connections.

Settings is a two-pane screen with a grouped rail on the left and one panel on the right. It holds what changes how the app behaves. What the agent can reach lives in [Connections](/openhuman/features/connections.md), and most of the old settings addresses for those now redirect there.

Open it with the gear in the sidebar header, or `⌘,`.

## The groups

### General

| Panel     | What it controls                                                                                           |
| --------- | ---------------------------------------------------------------------------------------------------------- |
| Account   | Who you are signed in as, your plan and usage, team, devices, notifications and language.                  |
| Privacy   | [Privacy mode](/openhuman/features/privacy-and-security/privacy-mode.md) and what may leave the machine.   |
| Migration | Moving your data folder, and importing from an earlier version.                                            |
| Core      | Point the app at a core running somewhere else: its RPC URL, a bearer token and a live reachability check. |
| Feedback  | Send feedback from inside the app.                                                                         |
| About     | Version, build, licenses and the capability catalog.                                                       |

### Appearance

| Panel        | What it controls                                                                                                                              |
| ------------ | --------------------------------------------------------------------------------------------------------------------------------------------- |
| Appearance   | Font size (including a custom pixel value), layout corners and borders, and language.                                                         |
| Theme Studio | Theme family and variant, per-token colors, per-role fonts, backdrop and your own saved themes. See [Themes](/openhuman/features/theming.md). |
| Personality  | The editable `SOUL.md` prompt and its companion identity file. See [Personalization](/openhuman/features/personalization.md).                 |
| Face         | Which mascot, its colors, a second "duo" mascot, a custom image and its voice. See [The mascot](/openhuman/features/mascot.md).               |

### Security

| Panel            | What it controls                                                                                                                                                                    |
| ---------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Keychain         | The OS keyring, consent, and what is encrypted at rest. See [OS keyring and secret storage](/openhuman/features/privacy-and-security/os-keyring-and-secret-storage.md).             |
| Agent access     | Trusted roots and their read or write grant, the always-allow list, the blanket auto-approve switch, action and tool timeouts, the workspace-only switch and the tool-call dialect. |
| Sandbox          | Which sandbox backend a session gets, and its limits.                                                                                                                               |
| Approval history | The audit trail of approval decisions and their outcomes.                                                                                                                           |

## Where the autonomy tier is

It is not in Agent access, which is where most people look first.

The tier (`readonly`, `supervised` or `full`) has no control in the normal settings navigation. The panel with the radios is developer-only and reachable only as a hidden deep link, so it does not appear in the sidebar. Set the tier in `config.toml`:

```toml
[autonomy]
enabled = true          # off by default; nothing below applies until this is true
level = "supervised"
```

`[autonomy] enabled` is `false` out of the box. Until you turn the policy on, the trusted roots, the always-allow list and the workspace-only switch in Agent access do nothing. [Approval gate](/openhuman/features/approval-gate.md) explains the whole model, including what stays enforced either way.

## Redirects

The panels moved into [Connections](/openhuman/features/connections.md), and the old addresses still work. Many more settings routes redirect than render. The main ones:

| Old address                                                                                                 | Now                                             |
| ----------------------------------------------------------------------------------------------------------- | ----------------------------------------------- |
| `/settings/llm`, `/settings/embeddings`, `/settings/voice`, `/settings/search`                              | Connections, the matching API-keys tab          |
| `/settings/agents/*`, `/settings/tools`                                                                     | Connections → Tools                             |
| `/settings/memory-*`, `/settings/intelligence`                                                              | Connections → Memory                            |
| `/settings/mcp-server`, `/settings/skills-runner`                                                           | Connections → MCP Servers, Connections → Skills |
| `/settings/wallet-balances`                                                                                 | Connections → Wallet Balances                   |
| `/settings/cron-jobs`                                                                                       | Workflows → Schedules                           |
| `/settings/automations`                                                                                     | Workflows                                       |
| `/settings/team`, `/settings/billing`, `/settings/notifications`, `/settings/devices`, `/settings/language` | Settings → Account                              |

## Not here yet

Settings has no search. Notification categories gate which events are ingested, but the page that edited them is gone, so they are config-only today.

## See also

* [Connections](/openhuman/features/connections.md): everything the agent plugs into.
* [Approval gate](/openhuman/features/approval-gate.md): the autonomy policy these panels configure.
* [Privacy and security](/openhuman/features/privacy-and-security.md): the posture behind the privacy panel.