DeepSeek puts its chat app's search, file reading and voice into the API
SummaryWeb search, deep thinking search, file input, URL reading, safety checks, OCR, speech to text and text to speech are now available to developers through the DeepSeek API, plus an open-source demo chat app. Not tested by MODEL.
· MODEL Editors · 2 min read
Some links may be affiliate links. Disclosure

What launched. On Friday, Nov. 13, @deepseek_ai announced (opens in a new tab) that "the capabilities behind the DeepSeek App are now available through the DeepSeek API." Per DeepSeek's announcement (opens in a new tab), developers can turn on the app's built-in features with a few parameters or call standalone APIs, and "recreate the full DeepSeek App" in their own product.
What it does. According to DeepSeek's chatbot guide (opens in a new tab), the V4.1-Flash model (deepseek-flash) now supports these features in the Messages, Responses and Chat Completions APIs, with tool calls run on DeepSeek's side:
- Web search: with thinking off and the
deepseek_web_searchtool on, the model searches the web or images and answers with citations. - Deep thinking search: with thinking on, the model searches and reads pages over several reasoning steps.
- File input: PDFs, Word, PowerPoint and image files are parsed into the conversation.
- URL reading: with
deepseek_read_urlon, the model fetches links in the user's message. - Safety checks: inputs and outputs are screened, and a refusal marker comes back when a risk is found.
- Search, fetch, OCR, speech to text and text to speech are also sold as standalone APIs.
DeepSeek also released an open-source demo chat client (opens in a new tab) (MIT license) that can import chat history exported from DeepSeek Chat. DeepSeek says the demo keeps your API key, chats and uploaded images in your browser, and that it is "for demonstration only."
What it costs. These tools are billed on top of normal token charges. DeepSeek's pricing page (opens in a new tab) lists, as of Nov. 13:
- Search (text or image) and web fetch: $0.75 per 1,000 uses each.
- File parsing and OCR: $0.225 per 1,000 pages.
- Text to speech: $0.45 per 10,000 characters, at a limited-time 50% off price of $0.225.
- Speech to text: $0.116 per hour of audio, at a limited-time 50% off price of $0.058.
DeepSeek says its prices may change.
Who it's for. Developers who want search with citations, file reading and voice in their own chat app without wiring up separate search, OCR and speech vendors.
The caveats.
- DeepSeek says public-facing apps must turn on safety checks, and developers are still responsible for content moderation, applicable model registration and AI-generated content labeling.
- The built-in tools run on
deepseek-flash. When they're on, DeepSeek's service sets some parameters itself, such as temperature, and ignores the values you send. - Every search, fetch and parsed page adds a per-use fee to your bill.
Disclosure: Not tested by MODEL. Not sponsored. No affiliate links.
MODEL Editors
Get three tested tools every Friday
Free, weekly. Unsubscribe anytime.
Sources
- The full DeepSeek App experience, now via API
DeepSeek API Docs
- Build a Chatbot
DeepSeek API Docs
- Models & Pricing
DeepSeek API Docs
- deepseek-ai/chat-demo
GitHub
- DeepSeek on X
X (@deepseek_ai)
Keep reading

Builder tip
AWS Emulator puts a local copy of AWS on your laptop. The builder tip: iterate there, test permissions on real AWS
AWS launched AWS Emulator on Nov. 11: a free local stand-in for its cloud that your CLI, SDKs and coding agents can call with no AWS account. AWS says AI agents wrote all of it. It doesn't evaluate IAM policies the way AWS does, so don't use it to check security. Not tested by MODEL.
· MODEL Editors

New in Beta
Claude Meetings puts Claude in your Google Meet call, in beta for Team and Enterprise
Anthropic's new beta adds Claude to a call as a visible participant that takes notes, answers questions and builds docs and whiteboards while you talk. Google Meet only for now. Not tested by MODEL.
· MODEL Editors

AI News
Gemini 4 Hydrogen is 2.5x faster than 3.8 Flash, and costs twice as much for now
Google's new everyday Gemini 4 model is rolling out in the Gemini app, Search, AI Studio and the Gemini API. It lists at $1.50/$7.50 per million tokens, double 3.8 Flash's promo price through Dec. 31. Not tested by MODEL.
· MODEL Editors

New Hardware
Google's Gemini glasses are up for pre-order: from $395, shipping Dec. 10
Google, Samsung, Warby Parker and Gentle Monster opened pre-orders for intelligent eyewear with Gemini built in: voice questions about what you see, Maps directions, live translation and a 12MP camera. Not tested by MODEL.
· MODEL Editors
FREE · EVERY WEEK
The top AI tools this week, tested.
Every week MODEL's editors pick the AI tools worth your time, try them hands-on, and tell you what's good, what's not, and who it's for. Free, in your inbox.
Free. Unsubscribe anytime. Our picks are editorial and can't be bought.