How to Use the OpenClaw TUI and Local Chat Mode

The OpenClaw TUI is the terminal interface for talking to an agent through the Gateway or through a local embedded runtime. It is useful when you want a focused operator console without jumping through Slack, Telegram, or a browser UI. The documented command is openclaw tui, with openclaw chat and openclaw terminal as aliases for local mode.

Gateway-backed mode

Use the default TUI when you want to connect to the Gateway. You can pass URL, token, password, session, and delivery options from scripts or runbooks. When launched from inside a configured agent workspace directory, the TUI auto-selects that agent for the session key default unless you explicitly pass an agent-scoped session.

openclaw tui
openclaw tui --url ws://127.0.0.1:18789 --token <token>
openclaw tui --session main --deliver
openclaw tui --session bugfix

Local embedded mode

Local mode uses the embedded agent runtime directly. Most local tools work, but Gateway-only features are unavailable. The docs are clear that --local cannot be combined with --url, --token, or --password.

openclaw chat
openclaw tui --local
openclaw chat --message "Compare my config to the docs and tell me what to fix"

Local mode also adds /auth [provider] inside the TUI command surface. Approval gates still apply, so tools that require approval prompt for a decision in the terminal. Nothing becomes silently auto-approved just because the Gateway is not involved.

Config repair loop

The docs recommend local mode when the current config already validates and you want the embedded agent to inspect it, compare it against docs, and help repair it from the same terminal. If config validation is already failing, use openclaw configure or openclaw doctor --fix first. Chat does not bypass the invalid-config guard.

openclaw chat

!openclaw config file
!openclaw docs gateway auth token secretref
!openclaw config validate
!openclaw doctor

When I reach for the TUI

• Testing an agent before exposing it to team channels.
• Repairing config with command output visible in one place.
• Using a named session for a focused investigation.
• Avoiding chat-channel noise while still using real agent tools.

The TUI is not glamorous, but it is dependable. In production ops, dependable beats glamorous almost every time.

If you are turning terminal-first OpenClaw operations into real operations instead of a demo, The OpenClaw Playbook is the shortcut I wish every operator had: identity files, memory rules, safety boundaries, channel discipline, and production habits in one field-tested guide.

Session labels and route hygiene

Named sessions are useful when the investigation has a theme, such as support triage or config repair. Do not reuse one label for unrelated work just because it is convenient. The session key carries context, and sloppy reuse can make the model blend old debugging assumptions into a new task.

Local mode boundaries

Local embedded mode is excellent when the gateway is not the thing you want to test. It is not a complete substitute for channel behavior, remote auth, pairing, or gateway-hosted routes. If a bug only appears in Slack, WhatsApp, or a remote node, reproduce it through the gateway path before declaring it fixed.

Runbook detail

For How to Use the OpenClaw TUI and Local Chat Mode, the important operator move is to record the exact documented surface you used and the condition that proves it worked. That might be a status command, a gateway event, a task record, a pairing approval, or a visible channel response. OpenClaw features are much easier to trust when the runbook says how to verify the feature, not just how to start it.

Operator checkpoint

Keep the first rollout narrow: one owner, one environment, one reversible test, and one written rollback note. Once the behavior matches the docs in that small setting, widen it deliberately. That habit prevents a useful OpenClaw feature from becoming another invisible system nobody knows how to debug. If the verification step is unclear, stop and tighten the runbook before you hand it to an autonomous agent.