DeepSeek Expands Developer API With Search, File and Voice Features

DeepSeek has expanded its developer API with tools for search, webpage retrieval, file analysis and voice, allowing developers to add these capabilities to products through its service.

Developers can enable tools in existing API requests. DeepSeek’s servers run the tool calls and combine the results with the model’s context. The model can search the web while reasoning, read linked pages and answer questions using uploaded PDFs, Office files and images. This setup handles the tool execution and result assembly that developers otherwise would need to build themselves.

The API also has separate interfaces for search, webpage retrieval, optical character recognition (OCR), speech recognition and text-to-speech. Developers can use the integrated chat features or select individual tools. DeepSeek had previously opened Search and Fetch API access to developers.

DeepSeek also released an open-source chat demo that supports multi-turn conversations, voice input, spoken replies and imports of DeepSeek chat histories. Developers can modify the demo to build their own DeepSeek chat product.