Draft:Naapi

NAAPI
Other names	naapi
Developers	NaCloud Network Technology Co., Ltd. (Taiwan, China)
Release	November 2025 (2025-11)
Platform	Web
Available in	Simplified Chinese, English, Français, Русский, 日本語, Tiếng Việt
Type	Large language model gateway / API forwarding service
Website	naapi.cc

NAAPI is a large language model (LLM) aggregation gateway and API forwarding service that follows the OpenAI API specification. It was launched in November 2025 and is maintained by Sodium Cloud Network Technology (Taiwan, China). The service provides a unified interface for accessing multiple LLM and multimodal AI endpoints from different providers.

Overview

NAAPI integrates APIs from several AI vendors, including OpenAI, Anthropic, Google, and Zhipu (Z.AI). It allows users to route all requests through a single base URL by changing only the API base URL and authentication key.

Key features include:

• Text and chat completions: supports the standard v1/chat/completions endpoint defined by OpenAI.
• Multimodal and image generation: also handles image generation and audio processing tasks.
• Data export and usage logs: users can export API call volumes, credit consumption, and task history.

Architecture and node distribution

NAAPI offers multiple routing paths optimized for different network environments:

• Main site and CDN routes: the primary domain naapi.cc, along with na.wanxiaot.com and 91vip.video, are hosted on Cloudflare for CDN acceleration.
• China-optimized nodes: specific nodes (e.g., toysheart.naapi.cc) are configured for users inside mainland China.
• Direct overseas nodes: nodes located outside China (e.g., us.naapi.cc, nanh.wanxiaot.com).
• Documentation site: API documentation is hosted on Apifox at naapi.apifox.cn.

Account and payment system

• Authentication: supports login via GitHub, Google, Discord, LinuxDO, Telegram, GitLab, passkeys, QQ, and Alipay, including OAuth and passwordless options.
• Top-up system: users purchase credits, which are then deducted based on actual usage.

Community monitoring data

According to third-party API aggregation monitoring platforms, as of July 2026, NAAPI listed approximately 376 model identifiers, with an average response time of around 1.90 seconds and a benchmark throughput of roughly 649 tokens per second.