Using the Astrenix AI gateway

One token reaches the major models, billed by actual usage from your account balance:

• Create a key: it starts with sk-nc and the plaintext is shown only once because the backend stores a hash; a monthly spend cap can be set
• An empty monthly cap means unlimited and 0 blocks all spend this month — the two are not the same
• Integration changes two things: point base_url at the gateway and swap api_key; the chat endpoint is /v1/chat/completions
• Marketplace prices are the official list price times the cheapest routing group; actual billing follows the group the request hits
• Usage and logs cover 24 hours, 30 days and 90 days with amount billed, total tokens, average latency and a per-call breakdown

Frequently Asked Questions

How much code has to change to integrate?

Usually just base_url and api_key. The gateway keeps official SDKs and protocols compatible, so request bodies, response structures and streaming behave the same.

Why is my bill higher than the marketplace price?

The marketplace price assumes the cheapest routing group, while billing follows the group actually hit. The exact ratio for any call is on its row under Usage and logs.

Related pages: Help Center · Guides & FAQ · AI API Gateway · Direct LLM Access · Tovanix API Documentation