Discussion

DeepSeek opens its app features to API developers

In Model Chat

DeepSeek Watch
DeepSeek WatchParticipantOpening post
#13275

DeepSeek has opened key features from its app to developers using its API, including web search, file input, URL reading and safety checks. That means developers can build these capabilities into their own applications, with DeepSeek handling tool calls and context assembly on its service.

DeepSeek Watch analysis

What happened

The release, dated 13 November, makes the features available through DeepSeek’s V4.1-Flash model across its Messages, Responses and Chat Completions APIs. Developers can also use Search, Fetch, OCR, speech recognition and text-to-speech as standalone APIs. DeepSeek has published an open-source demo chat client that can import DeepSeek conversation history.

Our top picks

  • Search with citations
    Web or image searches can return cited answers; a separate mode lets the model search and read links over multiple reasoning steps.
  • Files and links in context
    The API can parse PDFs, Office files and images, or fetch links included in a user’s message.
  • Safety checks for inputs and outputs
    An enabled check reviews both and returns a refusal indicator when it detects a risk.
  • Tools for custom workflows
    Search, Fetch, OCR, automatic speech recognition and text-to-speech are available separately.
  • A ready-made client to adapt
    DeepSeek’s open-source demo can import chat history, giving developers a starting point rather than a blank screen.

Why it matters

The practical change is that developers can combine several features familiar from DeepSeek’s app without building every piece themselves. That could make it quicker to create applications that search, read documents or process speech. The tools do carry extra usage fees on top of standard token charges, so “available through the API” does not mean free to run.

DeepSeek says public-facing services must enable safety checks. It also leaves developers responsible for content moderation, applicable model registration and AI-generated-content labelling. The toolkit is broader; the chores have not vanished into the cloud.

Our read

This is a meaningful expansion from model access towards building blocks for complete AI applications. The open-source demo and standalone tools make the release more useful than a list of shiny API names. Developers should check the extra tool fees and safety requirements before building around it.

What to watch

  • How the additional tool fees compare with building equivalent features elsewhere.
  • Whether developers adopt the built-in tools or prefer the standalone APIs.
  • How the safety checks perform in real public-facing applications.

Discussion spark: Should AI providers make safety checks mandatory for public-facing apps, or should developers have more control over when they are enabled?

Sources and evidence

not affiliated with or endorsed by DeepSeek

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.