New API is a unified AI model aggregation and distribution gateway that provides centralized model management for individuals and businesses. Its core positioning is not as a single model product, but as a “middle gateway” at the model-calling layer: it converts various large language models into OpenAI-, Claude-, and Gemini-compatible interfaces, reducing the complexity of adapting applications to different model APIs.
Based on the crawled text, New API focuses on model aggregation, model distribution, and cross-format interface conversion. For applications already built around the OpenAI, Claude, or Gemini protocols, it can serve as a unified entry point, allowing developers to centrally manage different model sources on the backend while minimizing changes to frontend or business logic. However, the text does not disclose key details such as which models are supported, whether it supports multimodal input, function calling, streaming output, rate limiting, logs, monitoring, key management, and other engineering-critical features. Its real-world production usability therefore still needs further verification.
The currently crawled content does not provide information on free quotas, trial policies, plan pricing, pay-as-you-go rules, or enterprise pricing, nor does it specify payment methods. If using it in a production environment, it is recommended to confirm the billing basis, how upstream model costs are passed through, concurrency limits, whether failed retries are charged, and whether a service-level agreement is available.
Its main advantage is clear positioning: by providing unified compatibility with OpenAI, Claude, and Gemini interfaces, it can reduce the cost of integrating and migrating across multiple models, making it suitable for teams that need model routing or centralized management. The limitations are also obvious: the publicly available information is repetitive and sparse, with little explanation of privacy and security, deployment options, stability, output quality, or support. Since it functions more like a gateway layer, the final output quality mainly depends on the underlying large language models being connected.
It is suitable for individual developers, AI application teams, and internal enterprise platform teams that want to wrap different large model APIs behind a unified interface. Access from China is not addressed in the text, so network connectivity, payment availability, and compliance support are all unknown. It may be compared with alternatives such as LiteLLM, OpenRouter, Portkey, Helicone, and One API, with the final choice depending on whether self-hosting, integration with Chinese domestic models, and enterprise audit capabilities are required.
⚠ This review is compiled from public sources and does not constitute a purchase recommendation. Verify all facts on the vendor's official site. Verify on 37it.com official site.
37it.com is an China AI Apps provider. TG4G tracks its product information, an overall rating of 7.0/10, and a China-accessibility score of China direct-connect friendly. Click "Visit Official Site" to reach 37it.com directly.