In the conversation service of the DeepSeek App (including the web version), capabilities such as web search, file input, link reading, voice interaction, and security work together with model inference to form the familiar chat experience for users.
If developers want to implement similar features in their applications, besides integrating the original DeepSeek API, they often need to spend a lot of effort connecting external tools and handling call integration.
Today, we officially open this full suite application API: by enabling the corresponding parameters in the original API, you can obtain chat features such as deep thinking search and file input; voice and related underlying capabilities have also been opened as independent interfaces (including OCR, ASR, TTS, etc.), supporting separate or combined calls.
Developers can easily implement complex chat features in their own applications, and can even recreate a complete DeepSeek App 1:1.
Fully opening application features, supporting independent calls
This update directly integrates the various interactive features of the DeepSeek App into the original API.
After enabling the corresponding parameters, related tool calls and context stitching are automatically completed by the server.
Currently, the V4.1-Flash model (deepseek-flash) already supports the following capabilities under the Messages, Responses, and Chat Completions interfaces:
Web search: In non-thinking mode, after enabling the deepseek_web_search tool, the model can automatically initiate web or image searches based on the question and answer with citations; Deep thinking search: In thinking mode, after enabling the deepseek_web_search tool, the model can autonomously search the web and read links during multi-step reasoning; File input: Supports uploading files in PDF, Office, image, and other formats, completing plain text extraction or mixed text-image parsing, then stitching them into the conversation context for the model to answer; Link reading: After enabling the deepseek_read_url tool, the model can automatically identify links in user messages, fetch webpage content, and answer in combination with context; Model security: After enabling security verification parameters, we will provide basic risk control, review model input and output, and return a refusal identifier when risks are hit.
If you provide application services externally, be sure to enable this parameter and assume obligations such as content security review, large model registration, and AI labeling.
In addition, to make it easier for developers to integrate their own workflows more flexibly, the above underlying capabilities and voice capabilities all support independent calls as separate interfaces, including: Search, Fetch, OCR, ASR, TTS.
Using the above interfaces, developers can implement functions basically consistent with the DeepSeek App.
To intuitively demonstrate this effect, we built and open-sourced a lightweight front-end Demo based on the new API, which supports directly importing DeepSeek historical conversations.
Demo experience entry: https://cdn.deepseek.com/chat-demo Sample code entry: https://github.com/deepseek-ai/chat-demo For more detailed parameter definitions and usage instructions, please refer to the DeepSeek official API documentation [API Guide - Building a Chatbot].
API pricing explanation
When users call the DeepSeek API to use extended capabilities, in addition to token fees, tool call fees will also be charged.
The prices of each tool are shown in the table below:
A few words at the end
We believe that interaction between humans and models is still full of infinite possibilities, and DeepSeek's existing chat product is only a relatively basic form.
Now, all capabilities behind the official product are supported for free calling and combination through the DeepSeek API.
We look forward to developers using this to save repetitive basic work and create more diverse, more refined, and more imaginative products.
Welcome to use the DeepSeek API and create freely.