Connectors

Connectors serve as a bridge connecting AI programming assistants to external services, data sources, or tools.

The status page marking only "connectors" in general terms means fault localization requires troubleshooting across multiple systems: the OpenAI inference layer, the connector gateway, the OAuth authorization layer and third-party service APIs.

When individual users encounter connector errors, they should first check the authorization status of the third-party service and the official status page; blindly deleting and reconnecting may trigger additional authorization process issues.

Recording connector call failure rates on the application side and cross-checking them against the official status page is more reliable than relying entirely on official notifications; regularly rotating OAuth tokens and monitoring token health can reduce non-server-side failures.

Teams that rely on Connectors to build internal tools should configure backup data sources or caching layers for critical connectors to implement degradation strategies, avoiding a single point of failure directly blocking business processes.

The OpenAI status page has had multiple Connectors-related incidents in the past with a highly similar pattern: it first shows "Investigating," updates to "Identified" or "Resolved" several hours later, and does not publish the list of specific affected connectors in between.

In agent scenarios, intermittent connector failures cause the model to retry, and each retry is a new model inference that consumes tokens and introduces latency, and the failure information may be misread by the model as "data does not exist" rather than "request failed".

The theoretical upper limit of a connector's availability equals the product of the availability of all dependencies in the chain; the more dependencies, the lower the overall reliability.

OpenAI Connectors itself does not directly promise an end-to-end SLA but inherits the SLAs of the third-party services it depends on.

When the OAuth refresh mechanism fails, a token is manually revoked by the user, or a third-party service changes its authorization scope requirements, the Connectors connector will immediately report an error.

OpenAI Connectors is an integration framework in products such as ChatGPT for connecting third-party applications (such as Google Drive, Slack, GitHub, etc.), allowing models to access external data sources and tools.

Luminous provides two integration paths through Connectors: a custom API and an MCP server.

Each connector provides fine-grained permission controls: Read-only, Needs Approval, Allow, and Block

Built-in Connectors include Google Drive, Canva, Slack, Gmail, Google Calendar, GitHub, and others

Cowork's Connectors obtain tokens for operating third-party services through an OAuth flow, which is the technical foundation for fine-grained permission control

Connectors allow Claude to directly interface with tools such as Google Calendar, Gmail, Google Drive, and Slack to perform actual operations, with the option to require approval for write operations and set automatic execution for read-only operations.

The newly added Browse area lists the full connector catalog in a two-column layout with search support, and the installed list and browse list each have their own independent search box.

In v0.0.42, Connectors becomes the first default tab in the Customize panel.

Cline's Connectors feature is essentially based on MCP or a similar mechanism, enabling the agent to invoke external capabilities such as file systems, browsers, databases, and third-party APIs.

When a connector is unavailable, a clear reason is provided: if the user is not logged in, a login guidance card is shown; if the user has not obtained beta access, a rollout prompt with a Check again button is displayed.