Using the gateway
For people who call models with a key, and for Team and Project admins.
These pages are for everyone with a platform role: people who create keys and call models from their code, and the owners and admins of Teams and Projects.
API keys
Create a key, choose its models and limits, and rotate, disable or revoke it.
Calling the API
The base URL, model names, and examples with the OpenAI and Anthropic SDKs.
Models
The models your workspace can use, their prices, and comparing them.
Logs
Requests, generations and sessions, with timing, tokens and cost. Never prompts.
Usage and costs
Spend, requests and tokens over time, broken down by model, key and member.
Alerts and notifications
Budget, spend, error-rate and batch alerts, in the dashboard and by email.
Files
Upload, list and download files with the OpenAI Files API.
Batches
Run thousands of requests as one job, for any model.
Realtime audio
Speech-to-speech sessions over a WebSocket.
Workspace settings
Members, invitations, service accounts, limits, access, alerts and the workspace's audit log.
The short version
- Pick a workspace in the sidebar: your Personal workspace, or a Team or Project you belong to.
- Open Models to see which models the workspace can use, and copy a model's name.
- Open API keys, choose Create key, and save the key somewhere safe. It is shown once.
- Point the OpenAI or Anthropic SDK at your gateway's address with that key, and use the model's name.
- Watch Logs and Usage & costs to see what your requests did and what they cost.
If a model you need isn't there, ask your workspace's admin (for a Team or Project) or a Platform Admin (for your personal workspace).