Token usage, costs and request volume per vKey (internal gateway usage)

Resale of shared providers: sales, settlement ledger and your own sharing configuration

Provider definitions and live usage of the upstream providers (/usages, /quotas)

Registered backend MCP servers: transport, health status, aggregated tools and cost per tool call

Registered A2A agents: agent card, protocol version and cost per call

Model groups with 3-level visibility (on / hidden / off)

Virtual keys: groups, limits, auth mode

Originals stored by the secret_redaction plugin (only your own). Values are hidden by default — click a value to reveal it.

Multi-tenancy: teams with members, roles and OIDC group mapping (provider ownership & sharing)

Request logs with TTFB, tokens, costs and ZDR status

Asynchronous batch processing (/v1/batch): status, progress, results — forwarded to provider batch endpoints or sync execution

Costs, throughput and TTFB per model — grouped by provider

All registered plugins: provider extensions, content filters, protocol translations and routers — enable/disable and configuration

WebSocket realtime endpoint (/v1/realtime) — currently a raw bidirectional proxy, not the full OpenAI Realtime protocol

System configuration: OIDC login (SSO), admin roles, gateway defaults (gateway admins only)