Introduction

Open Vertex Router is an AI gateway for teams. Point your existing OpenAI-compatible client at your gateway URL and reach models from OpenAI, Anthropic, Google and others through a single API key and a single monthly invoice.

Because the gateway speaks the OpenAI API, adopting it is a base URL change: no new SDK, no rewritten call sites. Spend is bounded by a prepaid balance, an optional credit line and per-key budgets, so a runaway job cannot outrun what you have approved.

What you get

• One integration. POST /v1/chat/completions and GET /v1/models, with streaming, exactly as your OpenAI client expects them.
• One key for every vendor. The model field picks the provider: gpt-4.1-mini, claude-haiku-4-5, gemini-2.5-flash. You never hold provider keys.
• One invoice, hard limits. Prepaid credits are consumed first, then an optional credit line with a monthly cap. Budgets per key and rate limits per plan bound the rest.
• Attribution. Tag each request by feature, environment or customer, and read spend by tag in the dashboard.
• Your own gateway. Each account runs on a dedicated gateway with its own URL, shown on the Integrate page of the dashboard.

Start here

QuickstartChange the base URL and the key, send a chat completion.

AuthenticationKeys, scopes, revocation, and what a rejected request looks like.

Making requestsEndpoints, streaming, and the few differences from OpenAI.

Credits and budgetsPrepaid balance, credit line, per-key budgets and alerts.