DeepSeek Opens Its App Tools to API Users

DeepSeek on Nov. 13 opened the tools behind its consumer chat app to developers, letting its API run web search, read links and uploaded files, screen content for safety, and turn speech into text and back. The Hangzhou-based AI lab announced the change in its API change log and in a thread on X, describing it as "the full DeepSeek App experience, now via API."

Key takeaways

• The deepseek-flash model can now search the web, read URLs and parse PDFs, Office files and images inside the Messages, Responses and Chat Completions APIs, with DeepSeek's servers running the tool calls.
• Search, Fetch, OCR, speech recognition (ASR) and text-to-speech (TTS) are also sold as standalone APIs, with tool fees charged on top of token prices.
• DeepSeek released an open-source demo chat client and says public-facing apps should turn on its safety checks.

What the DeepSeek API now does

According to DeepSeek's change log entry dated Nov. 13, developers switch the features on with parameters in existing requests. With thinking disabled and the deepseek_web_search tool enabled, the model runs web or image searches and answers with citations. With thinking enabled, the same tool lets it search and read pages across several reasoning steps.

A separate deepseek_read_url tool makes the model fetch links that appear in a user's message. File input accepts PDFs, Word, PowerPoint and image files, which are converted to text or parsed for layout before being added to the conversation.

The work happens on DeepSeek's side. Its Build a Chatbot guide says the service runs the searches and URL reading, and the client only shows the returned blocks. Until now, developers who wanted those features had to wire up their own search and parsing tools.

DeepSeek API pricing for search, OCR and voice

The tools are billed separately from tokens. DeepSeek's documentation lists these prices as of Nov. 13:

The voice discounts are labeled 50% off for a limited time, and DeepSeek gives no end date. Failed searches and empty results are not billed. The model itself, DeepSeek-V4.1-Flash-1017, costs $0.30 per million uncached input tokens and $1.20 per million output tokens at peak hours, with off-peak rates at half that, according to the pricing page.

Safety checks and the open-source demo

A safety-check parameter reviews model input and output and returns a refusal marker when it finds a risk. DeepSeek says developers serving the public should enable it and handle their own content review, large-model registration and AI labeling.

The company also published a demo client on GitHub. It runs in the browser, stores the API key and conversations locally, and can import chat history exported from DeepSeek's own app. DeepSeek calls it "for demonstration only."

Context: from model maker to tool platform

The launch builds on DeepSeek's Oct. 17 opening of its in-house search engine as a Search and Fetch API, and on the V4.1 Flash model it shipped in September, which AI Newsway covered at launch. Bundling search, document parsing and voice behind one model moves DeepSeek closer to the hosted-tool offerings of larger U.S. providers.

What's next

Two open questions are how long the voice discounts last and how the server-side tools perform at scale. AI Newsway did not test the APIs or the demo. TTS currently supports 10 languages, including English, Japanese and Korean, while ASR lists 21.

FAQ

Which DeepSeek model supports the new built-in tools?

The built-in tools work with deepseek-flash, which currently points to DeepSeek-V4.1-Flash-1017.

Do the tools cost extra?

Yes. Search, Fetch, OCR, ASR and TTS carry their own fees on top of token charges.

Is the demo app free to use?

The source code is public on GitHub, but it calls DeepSeek's paid API with the user's own key.

References

• Change Log — DeepSeek API Docs (Nov. 13, 2026)
• Build a Chatbot — DeepSeek API Docs
• Models & Pricing — DeepSeek API Docs
• Search & Fetch — DeepSeek API Docs
• OCR — DeepSeek API Docs
• Text to Speech — DeepSeek API Docs
• Speech to Text — DeepSeek API Docs