Ship AI features, not assembly.

opengateway is an AI platform for builders that lets you:

• Call supported frontier models through one OpenAI-compatible endpoint
• Switch providers by changing a string, not your code
• Read cost, latency, and logs for every request in one place
• Point coding agents at the same key your app already uses

You get one API key, one base URL, and one bill, because we connect directly to the companies building these models — there's no reseller sitting in the middle.

The 60-second test#

curl https://apis.opengateway.ai/v1/chat/completions \
  -H "Authorization: Bearer $OPENGATEWAY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-4o-mini",
    "messages": [{"role": "user", "content": "say hi"}]
  }'

Swap openai/gpt-4o-mini for anthropic/claude-sonnet-4 or google/gemini-2.5-pro and the request still works. The endpoint, the response shape, and your API key stay the same.

Why this exists#

A typical production AI stack looks like this:

App ─► OpenAI SDK ─► OpenAI
App ─► provider SDK ─► provider
App ─► a tracing tool ─► traces
App ─► a log service ─► logs
App ─► spreadsheet ─► "how much did we spend this week?"

Multiple SDKs, dashboards, billing surfaces, and places secrets can leak when someone pushes to GitHub by mistake. That is not a platform, it's assembly work.

opengateway collapses all of it behind one API surface so you can spend your time on the product you're actually trying to ship.

Start here#

• Quickstart — make your first API call in about 60 seconds
• Key Features — model namespacing, parameter normalization, debug mode
• Gateway — one endpoint, supported frontier models, OpenAI-compatible requests
• Observability — real-time cost, latency, and logs with no extra setup