OVTH Gateway onboarding in 5 minutes
One endpoint. Every model. Zero config. Drop-in OpenAI replacement.
⚡ 5min read beginner OS · macos · linux · windows v0.1.0
Last updated May 10, 2026 · by xlrd
Prereq
- OVTH API key from
ovth.dev/gateway - Any OpenAI SDK in any language
Steps
01. Set the base URL
bash
export OPENAI_API_KEY="ovth_live_..."
export OPENAI_BASE_URL="https://gateway.ovth.dev/v1"02. Test with curl
bash
curl $OPENAI_BASE_URL/chat/completions \
-H "Authorization: Bearer $OPENAI_API_KEY" \
-d '{
"model": "claude-opus-4.7",
"messages": [{"role":"user","content":"hello ovth"}]
}'03. Try different models
bash
# Claude (Anthropic)
curl ... -d '{"model":"claude-opus-4.7", ...}'
# Gemini (Google)
curl ... -d '{"model":"gemini-3-flash", ...}'
# Kimi (Moonshot)
curl ... -d '{"model":"kimi-k2.6", ...}'Next
- See all supported models at
/v1/models - Check prices, context windows, reasoning, vision, and tool support at
/v1/limits - For images, send uploads as base64/data URLs or pass a public image URL reachable from the gateway
Feedback · anonymous
Was this helpful?