# AIHubMix A unified API gateway for developers building with models from multiple AI providers. - Findings24: - Website: - Pricing: Freemium - Categories: AI Gateway And Routing, AI Model Hubs, LLM API ## What is AIHubMix? AIHubMix is an API gateway and model marketplace for accessing AI models from multiple providers through a common service. It offers OpenAI-compatible endpoints alongside protocol-specific interfaces, allowing developers to connect supported models using familiar SDKs and API patterns. A public catalog lists model capabilities and prices, while a console provides self-service API key creation and controls for restricting keys by model, IP or CIDR range, spending cap, and expiry date. Requests are billed at the selected or automatically routed model's listed rate; the automatic router itself is described as free and in beta. ## What can AIHubMix do? ### Unified model catalog Browse a catalog spanning providers including OpenAI, Anthropic, Google, DeepSeek, Qwen, and others, with model identifiers, supported modalities, and pricing information. ### Multiple API protocols Use Chat Completions or Responses with bearer authentication, Messages with an x-api-key header, or Gemini endpoints with x-goog-api-key, according to the model and SDK. ### Self-service key controls Create API keys in the console and configure allowed models, IP or CIDR allowlists, spending caps, and expiration dates. ### Automatic model routing The beta LLM Router accepts the model name auto and selects a model based on request content. The routed model's normal price applies; routing has no separate charge. ### Interactive model testing Use the Playground to try a model before integrating it. Free models on the production API base URL are the documented test path; there is no separate sandbox host. ### Usage-aware pricing The catalog displays per-model rates, including token-based prices and, for some capabilities, request-based charges. Review the selected model's pricing before sending paid requests. ## Use Cases ### Integrate models into an application Connect an existing SDK or service to AIHubMix, select a catalog model ID, and send requests using the protocol that model supports. ### Compare models before adoption Use catalog specifications and pricing, then test candidates in the Playground or through the supported free-model path before choosing a model for a workload. ### Control API key exposure Issue a restricted key for a service or environment, limiting the models it can call and optionally setting network allowlists, spend caps, and expiry dates. ### Route varied requests automatically Test the beta auto router when an application handles requests with different capability needs and would benefit from selecting a model based on request content. ## Frequently Asked Questions ### How do I make a first API request? Create an API key in the console, select a real model ID from the catalog, and use that model's documented protocol and integration guide. Chat Completions and Responses use Authorization: Bearer; Messages and Gemini use their respective authentication headers. ### Can I test without setting up a separate sandbox? The documented test options are the Playground and free models on the production API base URL. AIHubMix does not provide a separate sandbox host or mock API. ### How does the auto router affect cost? The beta router dispatches requests to a model based on their content. The routed model's price is charged, and the router itself has no separate fee. ### How can I limit what an API key can do? The console lets you restrict a key to an allowed model list and configure an IP or CIDR allowlist, a spending cap, and an expiry date. ### Where can I check model prices? Prices and model details are listed in the public catalog. Charges vary by model and may be based on tokens or other listed units, so review the selected model's entry before making paid requests. ## Quick Facts - **Product type**: AI model API gateway and marketplace - **API base domain**: aihubmix.com - **Model access**: Catalog includes models from multiple AI providers - **Integration**: OpenAI-compatible endpoints and other model-specific protocols - **Billing**: Per-model pricing; rates and billing units are shown in the catalog - **Router status**: Automatic LLM Router is in beta