Locations
LivRouter serves every model from one global endpoint: https://api.livrouter.com/v1. There are no regional endpoints to pick between and no per-provider base URLs to maintain — the model parameter in each request decides which model serves it, not the URL.
Which base URL to use
- OpenAI-compatible clients — OpenAI SDK, SillyTavern, OpenCode, and custom apps — use
https://api.livrouter.com/v1. - Anthropic SDK and Claude Code use
https://api.livrouter.com, without the/v1suffix: the SDK appends/v1/messagesitself. Pointing these clients at the/v1URL is the most common cause of 404 responses.
Check connectivity from your environment
Before wiring up an SDK, confirm that your machine reaches the endpoint and that your key works. The request below prints the HTTP status and total time:
curl -sS -o /dev/null -w "%{http_code} %{time_total}s\n" \ https://api.livrouter.com/v1/models \ -H "Authorization: Bearer $OPENAI_API_KEY"A 200 means the endpoint and key are good. A 401 means the key is wrong or missing; create one on the API keys page. A 404 usually means the path is wrong — re-check the /v1 rules above. Since requests are served through one endpoint, compare response times from your own environment when latency matters for your integration.
What is shared and what is not
- API keys, prepaid credits, and usage logs are account-level: one key works for every model, and all usage lands in the same dashboard.
- Model availability and pricing follow the live catalog — check the model catalog before pinning a model ID.