What changed
The full DeepSeek App experience, including search, files, links, safety, and voice capabilities, is now available through the DeepSeek API. Developers can use built-in features with a few parameters or call standalone APIs such as OCR, ASR, and TTS. The V4.1-Flash (deepseek-flash) model now supports these capabilities across Messages, Responses, and Chat Completions, with tool calls and context assembly handled server-side.
What this means in practice
Developers can build richer conversations or recreate the full DeepSeek App in their own applications. The API includes features like Web Search, Deep Thinking Search, File Input, URL Reading, and Safety Checks. These tools allow developers to search the web or images with citations, parse PDFs and Office files, automatically fetch links in user messages, and screen inputs and outputs for risks. Public-facing apps must enable safety checks, and developers remain responsible for content moderation, model registration, and AI-generated content labeling. Standalone APIs for search, fetch, OCR, ASR, and TTS are also available for custom workflows.
Limitations and availability
The API pricing for extended capabilities includes tool usage fees in addition to standard token charges. Developers can try an open-source frontend demo and access code on GitHub. The DeepSeek App is one way to interact with AI, but its capabilities are now available for developers to call and combine freely. The supplied announcement does not specify exact pricing details or the availability of all features in different regions.
