home-assistant/core · error · UpdateFailed

api_error

Error message

api_error

What it means

An UpdateFailed raised with translation key 'api_error' when the Anthropic coordinator's models.list() call raises a generic anthropic.APIError (anything other than timeout or authentication: 400/403/429/5xx). The DataUpdateCoordinator catches it, marks the entity unavailable, and schedules the next poll.

Source

Thrown at homeassistant/components/anthropic/coordinator.py:99

        self.update_interval = UPDATE_INTERVAL_CONNECTED
        super().async_set_updated_data(data)

    async def async_update_data(self) -> list[anthropic.types.ModelInfo]:
        """Fetch data from the API."""
        try:
            self.update_interval = UPDATE_INTERVAL_DISCONNECTED
            result = await self.client.models.list(timeout=10.0)
            self.update_interval = UPDATE_INTERVAL_CONNECTED
        except anthropic.APITimeoutError as err:
            raise TimeoutError(err.message or str(err)) from err
        except anthropic.AuthenticationError as err:
            raise ConfigEntryAuthFailed(
                translation_domain=DOMAIN,
                translation_key="api_authentication_error",
                translation_placeholders={"message": err.message},
            ) from err
        except anthropic.APIError as err:
            raise UpdateFailed(
                translation_domain=DOMAIN,
                translation_key="api_error",
                translation_placeholders={"message": err.message},
            ) from err
        return result.data

    def mark_connection_error(self) -> None:
        """Mark the connection as having an error and reschedule background check."""
        self.update_interval = UPDATE_INTERVAL_DISCONNECTED
        if self.last_update_success:
            self.last_update_success = False
            self.async_update_listeners()
            if self._listeners and not self.hass.is_stopping:
                self._schedule_refresh()

    @callback
    def get_model_info(self, model_id: str) -> tuple[anthropic.types.ModelInfo, bool]:
        """Get model info for a given model ID."""

View on GitHub (pinned to 58a3fdb3ea)

Solutions

1. Check err.status_code via logs to distinguish 429 (slow down usage / raise rate limits) from 5xx (wait for recovery)
2. Upgrade the anthropic package to a version pinned by Home Assistant core requirements
3. If persistent 4xx, inspect the message placeholder for the API's explanation (bad model id, org issues)

try:
    result = await self.client.models.list(timeout=10.0)
except anthropic.APITimeoutError as err:
    raise TimeoutError(err.message or str(err)) from err
except anthropic.AuthenticationError as err:
    raise ConfigEntryAuthFailed(...) from err
except anthropic.APIError as err:
    raise UpdateFailed(...) from err

When it happens

Trigger: await self.client.models.list(timeout=10.0) raising anthropic.APIError: HTTP 429 rate limit, 500/529 overloaded Anthropic service, 400 bad request from a library/model mismatch, or 403 permission errors not classified as AuthenticationError.

Common situations: Rate limiting after heavy conversation agent usage, Anthropic capacity events (529 overloaded), or an anthropic package version whose request shape no longer matches the API.

Related errors

• err
• api_authentication_error
• Error communicating with API: {err}
• Error communicating with API: {err}
• Unable find multi-factor auth module: {mfa_module_id}

AI-assisted analysis of home-assistant/core@58a3fdb3ea (2026-08-14). Data as JSON: /api/errors/3c07c0f8ddc45679. Report an issue: GitHub.