DeepSeek opens full-suite app API for replicating its consumer app

On the evening of November 13, 2026, DeepSeek announced the official opening of its full-suite application API. The update folds the interaction features of the DeepSeek App directly into the existing API, so developers can enable the relevant parameters and gain conversational capabilities such as deep-thinking search and file input, with tool invocation and context assembly handled automatically on the server side. DeepSeek says this makes it possible to reproduce the complete DeepSeek App one-to-one within third-party applications.

According to the official announcement, voice and related low-level capabilities have also been exposed through separate interfaces, covering OCR, ASR and TTS, and can be called either individually or in combination. At present, the V4.1-Flash model (deepseek-flash) supports these capabilities across the Messages, Responses and Chat Completions interfaces. With the deepseek_web_search tool enabled in non-thinking mode, the model can automatically run web or image searches based on a query and answer with citations attached. In thinking mode, the same tool allows the model to search the web and read links on its own while carrying out multi-step reasoning. File input supports uploads of PDF, Office and image files, together with plain-text extraction and similar processing.