Naver is revising its terms of service to prevent external developers from using its open search API for artificial intelligence (AI) training and answer generation. The company is also migrating the search API distribution channel from its existing developer center to the “Naver API Hub” on the Naver Cloud Platform (NCP) and introducing usage-based billing.

The move effectively seals off channels through which Naver’s search data could flow to external AI systems, timed to coincide with the expansion of AI Tab, its conversational AI search service.

According to Naver on the 24th, the Naver developer center recently announced revisions to the “Naver API Service Terms of Use.” The current terms only stipulate that search results must not be “independently exposed or distorted.” The revised terms explicitly prohibit: ▲ inputting data into AI systems or using it for training, improvement, evaluation, or exposure; ▲ profiting from use of the search API; ▲ providing or selling data to third parties; and ▲ copying, storing, or caching data. The revision takes effect on the 7th of next month.

An API allows external programs to receive specific service functions and data in a defined format. Developers can use Naver’s open search API to display Naver search results in their own web or mobile apps across categories such as news, encyclopedias, blogs, shopping, web documents, professional information, Knowledge iN, books, and cafe posts. Naver opened its search service as an open API in 2006 to expand its ecosystem.

The core of the revision is the “AI use ban.” It restricts not only training AI models with Naver search data, but also inputting data into AI systems and exposing results. Blocking training alone would leave room for retrieval-augmented generation (RAG), which pulls search data in real time to display in answers, but prohibiting exposure effectively closes the door on external AI using Naver’s search API. The ban on resale and monetization adds another layer of restriction.

Naver has been blocking unauthorized AI data collection through crawling since last year, and this API terms revision extends restrictions to AI use of data provided through official APIs.

Separately, Naver recently changed the entity providing three APIs—search API, search term trends, and shopping insights—from the developer center to the Naver API Hub. The search API was previously free on the developer center, but the Naver API Hub applies usage-based billing for calls exceeding the free quota. The search API has effectively transitioned to a paid service. However, Naver has not yet implemented the fees and announced they will be applied at a later date.

Aimed at Protecting AI Tab’s Competitiveness

The move is widely interpreted as being aligned with Naver’s AI search strategy. Naver officially launched “AI Tab” for all users in June. It provides answers based on the company’s own content, including blogs, cafes, shopping, and places. The service aims to be an action-oriented AI that leads to purchases and reservations.

AI Tab’s competitiveness stems from Naver’s content and Korean-language data. If external AI systems use the same data through Naver’s search API, AI Tab’s competitive edge could weaken. Naver decided starting on the 1st of this month to stop providing shopping search data through its API, as shopping data is a core asset that enables AI Tab’s action-oriented search.

There are precedents internationally of companies closing free search APIs and moving to paid models. Microsoft (MSFT) terminated its Bing Search API in August 2025, which developers had used to add search functionality to their services. The company instead introduced “Grounding with Bing Search” within its Azure cloud as an alternative, but using this service requires building an Azure project and paying per call.

Lee Kyung-jun, a professor in the Department of Big Data Applications at Kyung Hee University, analyzed: “Cases have emerged where various AI services call Naver’s API to use Naver search data for training, incorporate it into their own services, or where AI agents directly call the API.” He added, “This is a measure to prevent that.”

He explained: “Going forward, the content that AI utilizes will become as important as model performance in AI services. Naver sees the power of its AI services as coming from the content the company holds, and is trying to prevent potential competitors from indiscriminately accessing its API.”