Using Smart Search
Search combines answers with verifiable pages from the web. Questions about current versions, news or prices can therefore use recent information instead of only stored knowledge.
Use DeepSeek for conversation, writing, coding and working with documents. This page brings together its web, mobile, Harness desktop and API entry points, with an explanation of reasoning and search.
DeepSeek brings conversation, multi-step reasoning, verifiable web search and coding into one assistant. Those capabilities are accessible through four product entry points.
The chat interface provides Deep Thinking and Smart Search switches. The first expands reasoning; the second uses verifiable web information in the answer.
Search combines answers with verifiable pages from the web. Questions about current versions, news or prices can therefore use recent information instead of only stored knowledge.
Some DeepSeek models have openly released weights. R1 uses an MIT licence and supports research, local deployment and fine-tuning.
Harness is a desktop agent for documents, spreadsheets and code. Plugins extend its capabilities, and longer tasks can continue in the background.
Your application can call the models through a request format compatible with mainstream SDKs. Streaming responses and structured output are supported.
Use it to draft and revise text, summarise long reports, compare options and organise raw material into structured conclusions.
The model works through its reasoning before giving an answer. This is suited to mathematics, logic and tasks requiring several steps.
All four access options use the same underlying models. They differ mainly in use scenario and whether installation is required.
| Access method | Supported devices | Installation required? | Open the service |
|---|---|---|---|
| DeepSeek App | iOS or Android devices | Yes; use the official app store | Find the official DeepSeek app in App Store or a mainstream Android store |
| DeepSeek Harness desktop | Desktop computers | Desktop app installation required | Visit the official Harness page |
| DeepSeek API access | Applications and servers | No installation; register and create an API key | platform.deepseek.com |
| DeepSeek browser access | Devices with a web browser | Not required | chat.deepseek.com |
DeepSeek is a general-purpose AI assistant and model family developed by Hangzhou DeepSeek. DeepSeek is the brand spelling; the spaced English form Deep Seek refers to the same product.
The web chat can be used free of charge. API calls are billed for input and output tokens; rates depend on the model and the platform pricing page.
The web version needs no download: open the official chat address in a browser. For mobile, look for the official app in App Store or a mainstream Android store; the official Harness page provides desktop access.
Yes. Smart Search uses verifiable web sources in the browser app. The API also offers web search and page crawling, documented in the official API guide.
The Chinese pages provide more detail on models, API setup and practical tutorials.
Compare V4.1, V4 and R1 against the task you need to complete.
Read the model comparisonTry runnable curl, Python and JavaScript examples, with guidance on errors.
Open the API integration guideSix practical tutorials cover prompts, coding, documents, local deployment and API engineering.
Browse the tutorials