AI providers

Provider health and model routing for the AI pipeline. Platform staff only — customers never see provider or model controls.

Provider overview

Live connectivity checks. Credentials are reported as “Configured” or “Not configured” only — no secret values are stored or displayed here.

Provider	Configured	Status	Model	Latency	Note
Supabase	Not configured	Not configured	—	—	Not configured
Gemini (research)	Configured	Failed	`gemini-1.5-flash`	249 ms	HTTP 404
Claude (risk analysis)	Configured	Healthy	`claude-sonnet-4-6`	965 ms	Connected
OpenAI (communication)	Configured	Healthy	`gpt-4o-mini`	705 ms	Connected
Google Places (discovery)	Configured	Healthy	—	233 ms	Connected
Background jobs (Trigger.dev)	Configured	Healthy	—	—	Configured (inline fallback otherwise)

Model routing

The active model for each stage of the pipeline.

Research (Gemini)

Risk complexity (Claude)

Communication (OpenAI)

Defaults are set via environment; per-agency overrides are managed here in a future release.

Provider and model configuration is internal to platform staff. Customers never see provider or model controls, and no secret values are rendered anywhere on this page.