AI Gateway / MCP Runtime

Pull request logs from an LLM gateway or MCP server (LiteLLM, Portkey, Helicone, or an in-house proxy) to see which agents call which tools and models, on whose behalf. Covers every app behind the gateway without instrumenting each one. For apps you control, the aisoc-ai-sdk push path is lower friction.

• Connector id: ai_gateway
• Category: ai

Configuration

Field	Label	Type	Required	Default	Notes
base_url	Gateway base URL	string	yes	—	—
api_key	Gateway API key	secret	yes	—	encrypted at rest
approved_models	Approved models (comma-separated)	string	no	—	Any model outside this list raises an unapproved_model finding. Leave blank to disable the check rather than flag everything.
verify_tls	Verify TLS certificate	boolean	no	True	Disable only for an internal gateway using a private CA.

Credential handling

api_key is encrypted at rest by the credential vault before the row is written, and decrypted only at poll time. The value is never logged and never returned by the API — a read of the saved connector shows the field as set, not its contents. See credential vault.

Capabilities

• pull_alerts
• pull_audit

Setup

1. In the console, go to Connectors → Add connector and pick AI Gateway / MCP Runtime.
2. Fill in the fields above and run Test connection. The test is read-only and does not save anything.
3. Save. Polling starts on a jittered schedule within about a minute; the connector row shows last_sync once the first poll completes.

If last_sync stays empty

The first poll is the honest signal that the credential works. A connector that saves cleanly and never syncs usually means the credential lacks a scope the vendor requires for reads — check the connectors service log for the vendor's own error, which is passed through verbatim rather than summarised.

Vendor-specific setup notes go here: which console page issues the credential, which scopes to grant, rate limits, and anything the schema cannot express. This section is preserved when the page is regenerated.