DeepSeek has launched a new search and fetch API, priced at $0.75 per 1,000 successful requests, marking the Chinese AI company's latest move to expand its developer tools ecosystem. The API allows developers to integrate web search capabilities directly into their AI applications, enabling retrieval-augmented generation and real-time information access.
The pricing undercuts many existing search API providers, positioning DeepSeek as a cost-effective option for developers building AI-powered search features. The service targets applications that need to combine large language model capabilities with up-to-date web information, a pattern increasingly common in chatbots, research assistants, and enterprise AI tools.
The launch comes as competition intensifies among AI infrastructure providers. Companies like OpenAI, Anthropic, and Google have been expanding their API offerings beyond core language models to include tools for search, code execution, and data retrieval. DeepSeek's entry adds another player to this growing market, with pricing that could pressure incumbents.
The search API supports standard REST endpoints and returns structured results optimized for LLM consumption. Developers can use it to give their AI models access to current information without building custom web scraping infrastructure. The service includes rate limiting and usage tracking through a standard dashboard.
For developers in the DeepSeek ecosystem, the API provides a native way to add search capabilities without switching providers. The company has been gaining traction with its open-weight models and competitive pricing, and this expansion into tooling suggests a broader platform play beyond just model inference.