LLM Gateway

⚡ LLM Gateway
Dashboard Users Batches Models Logs Settings

Overview

Daily spend (USD)

Daily requests / active users

Top users

By model

By batch

⬇ Export CSV
0 selected

Batches

A batch is a group of users. Batch limits apply to every member unless a user has their own override.

Model access

Users can only call enabled models. Disabled ones are hidden from /v1/models (so they vanish from Claude Desktop's picker) and requests to them get 403. Sync adds newly released models as disabled.

Recent requests

Pricing (USD per 1M tokens)

Model id prefix match (longest wins). default is used when nothing matches. Fields: input, output, cache_read, cache_write_5min, cache_write_1hour. Used to compute cost + enforce $ limits.

Claude Desktop setup

Connection: Gateway · Credential kind: Static API key
Gateway base URL:
Gateway API key: the user's key from the Users tab
Auth scheme: x-api-key or Bearer — both work