Provider authentication

Happier can inspect and manage the machine-local authentication state of supported provider CLIs directly from each provider settings page.

This is separate from Connected services:

• Provider authentication = the native login state of the provider CLI installed on a specific machine
• Connected services = credentials stored in Happier and reused across compatible backends

This also differs from Custom ACP:

• built-in providers such as Claude, Codex, Gemini, or Kiro have dedicated provider pages
• Custom ACP stores readiness/login metadata inside each ACP backend definition instead of exposing one single provider-wide auth state for every custom backend

Where to find it

Open:

1. Settings
2. AI provider settings
3. Choose a provider
4. Choose the target machine in the context bar
5. Look for the Authentication section

The Authentication section is machine-specific. If you switch to a different machine, Happier checks that machine’s provider CLI instead.

What Happier shows

Depending on the provider, Happier can show:

• Status — Logged in, logged out, or unknown
• Logged in as — an account email or label when the provider exposes one safely
• Method — for example API key, auth token, OAuth CLI login, credentials file, or Google ADC
• Source — whether the result came from environment variables, local files, or a provider CLI command
• Issue — why login could not be confirmed, such as missing credentials or expired credentials
• Last checked — when Happier last refreshed that auth state

Actions

Log in / Re-authenticate

If the provider supports a native CLI login flow, Happier can open it directly inside the app.

What happens:

• on larger layouts, the login terminal opens in the same shared bottom-pane system used by session terminals
• on phone-sized layouts, Happier uses a more focused terminal presentation
• Happier launches the provider’s real CLI on the selected machine
• if the provider needs an in-terminal command such as /login, /auth, /setup, or /connect, Happier sends it for you
• if the provider prints a URL, Happier shows copy/open actions in the terminal UI

When the login flow exits, Happier refreshes the provider auth state automatically.

Check now

Check now forces an immediate auth refresh for the selected provider on the selected machine.

Use it when:

• you logged in outside Happier and want the status to update immediately
• you changed environment variables
• you want to re-check a previously unknown status

What “unknown” means

Some providers expose a safe, non-interactive way to confirm local auth. Others do not.

If Happier shows Unknown, that usually means one of these:

• the provider does not expose a reliable non-interactive auth probe yet
• the provider CLI check timed out or failed
• Happier intentionally avoided an interactive provider command during background detection

Unknown does not automatically mean “logged out”.

Which providers support what

“Limited / conservative” means Happier can launch the login flow, but only reports a positive auth state when the provider exposes a safe local signal.

How this works with Custom ACP

Custom ACP is more generic than a dedicated provider page.

For Custom ACP:

• auth/login hints live on the ACP backend definition

• different ACP backends can expose different login/status behavior So if you are using Custom ACP, think in terms of:

• “is this backend ready on this machine?”

• not “does the generic Custom ACP provider have one global login state?”

See ACP backends.

How the login terminal fits with the rest of Happier

The provider login terminal is not a one-off modal. It reuses the same terminal foundation as Happier’s embedded session terminal:

• the same connected-machine terminal transport
• the same URL detection and open/copy actions
• the same bottom-pane host on larger layouts
• the same close / restart / clear behaviors

That keeps provider login flows consistent with the rest of the app instead of introducing a second terminal UI.

Common use cases

“I installed a provider CLI and want to finish setup in the app”

Open the provider’s settings page, confirm the CLI is detected, then use Log in.

“I already logged in in my terminal and want Happier to pick it up”

Open the provider’s settings page and use Check now.

“I want to know which account the machine is using”

If the provider exposes an account label safely, Happier shows it in Logged in as.

“I use env-based auth instead of interactive login”

Happier can detect env-backed auth for providers that expose it and shows the auth method accordingly.

Important limitations

• Provider authentication is machine-local. Logging into a provider on one machine does not log it in on another machine.
• Not every provider supports reliable non-interactive auth detection.
• Connected services and machine-local auth can both exist for the same provider. They solve different problems.
• If a provider login flow changes upstream, Happier follows the provider CLI’s real current behavior rather than inventing a custom auth flow.
• For Custom ACP, authentication/readiness is backend-specific rather than one global provider-wide state.

Provider authentication and session handoff

This matters directly for session handoff between machines.

When you move a session to another machine:

• machine-local provider auth does not come along automatically
• the target machine must still be able to run that provider/backend

If the provider/backend supports Happier Connected services, that is often the easiest way to make the target machine ready for handoff without depending only on a native CLI login on that machine.

If the provider/backend depends on native machine-local auth, you still need the destination machine to be authenticated.

Related guides

• Providers
• ACP backends
• Session handoff
• Connected services
• Embedded terminal
• Daemon