NomiFun/05 MODEL ACCESS FOR YOUR BUSINESS
Run a customizable token billing service for NomiFun communities, teams, and customers. Self-host a model gateway to bring upstream model access, user management, and billing into your own service.
Operators deploy and price their own services. Users choose a service to connect to, with model calls, usage, and accounting managed in one gateway.
BUILT AROUND YOU
An independent model gateway released under Apache-2.0. Choose providers, configure models and prices, and manage your service through an English/Chinese web console. Users connect through NomiFun Desktop or native API clients.
Connect OpenAI, Anthropic, Gemini, OpenAI-compatible, and Azure OpenAI channels according to each model's supported capabilities. Publish a model catalog and token prices using provider templates, catalog discovery, or manual configuration.
Manage users and API keys in the embedded console, with model scopes, budgets, rate limits, and concurrency limits. Users can view models, prices, accounts, and usage, then connect Desktop with the instance URL and their key.
Manage wallets, subscriptions, redemption codes, and usage settlement with a traceable ledger. Stripe, Alipay, and WeChat Pay v3 merchant adapters are available for operators to configure and validate before enabling payments.
Customize operator details, purchase links, and service prices. A Go service embeds the web console, supports SQLite or PostgreSQL, and includes Docker deployment and integration guides for running an independent service.
Operators self-host Model Gateway and are responsible for pricing, terms, privacy, payments, and user support. NomiFun provides open-source software and does not operate gateway instances or sell API access. Desktop integration is optional and requires a version with the Model Gateway provider. Payments require your own configured and validated merchant accounts. Requests stay in the same native protocol; one database supports one running gateway process. See the repository for supported models and deployment requirements.
Read the repository guide