Fadhel Hasbullah
← All work

OpenXense

Co-founder & co-owner · Built from zero · 2025–now

Visit openxense.com ↗

One product for AI chat and developer APIs. I built the gateway, the billing engine and the SaaS experience, and I operate its infrastructure.

Product architecture · request flow

Chat + APIUsers and developers
Go gatewayAuth · keys · quota
LiteLLMRouting · max_budget
AI providers

usage

BillingQR-code payments
Built: gateway, billing, dashboard, chat. Used: LiteLLM, model providers.

The problem

Paying every AI provider separately adds up. People want chat; developers want an API. Both want one bill.

Why our own gateway

We could have exposed LiteLLM directly. A Go gateway in front of it keeps authentication, keys, quotas and traffic tracking in code we own.

Prepaid billing

Quota is denominated in USD, checkout is in IDR with QR-code payments, and margins are enforced through LiteLLM's max_budget per user.

See the product

Stack: Go, LiteLLM, PostgreSQL, Redis, gRPC, Docker, CI/CD.