// how to
How to use DeepSeek
Pick what you want to do. Each guide gives the steps, prompts you can copy into DeepSeek, an example of a good result, and the mistakes that waste your time.
DeepSeek how-tos
- 01
How to Use DeepSeek: A Beginner's Guide to the Chat App
How to use DeepSeek for the first time: sign up, ask clearly, switch on DeepThink or Search when they help, and know where your data is stored.
- 02
How to Run DeepSeek Locally: Which Model, What Hardware, Which App
How to run DeepSeek on your own computer: pick a distilled DeepSeek-R1 size that fits your memory, install Ollama or LM Studio, and know the limits.
- 03
How to Use DeepThink in DeepSeek (and When to Leave It Off)
How to use DeepSeek's DeepThink reasoning mode: when it helps (math, logic, code, plans), when it slows you down, and prompts that suit it.
- 04
How to Use the DeepSeek API: Key, Prices and First Call
How to use the DeepSeek API: create a key, top up, call deepseek-flash with the OpenAI or Anthropic SDK, and understand peak and off-peak prices.
- 05
How to Keep Your Data Private in DeepSeek: Settings and Limits
How to keep your data private in DeepSeek: what it collects and stores in China, how to opt out of training, delete chats, and when to run it locally.
- 06
How to Fix DeepSeek's "Server Is Busy" Message
What to do when DeepSeek says the server is busy: check status.deepseek.com, wait and retry, rule out your own connection, and keep a backup ready.
- 07
How to Code With DeepSeek: Chat, Coding Agents and Harness
How to code with DeepSeek: ask in the chat with DeepThink on, connect coding agents like Claude Code or OpenCode to its API, or try DeepSeek Harness.
- 08
How to Run DeepSeek With Ollama, Step by Step
How to run DeepSeek-R1 with Ollama: install it, run one command, choose a size from 1.5b to 70b, use the local API, and avoid cloud models by mistake.
- 09
How to Upload Files and Images to DeepSeek (and Fix Failed Uploads)
How to upload documents and images to DeepSeek, ask useful questions about them, fix failed uploads, and know which files never to upload.
- 10
How to Use DeepSeek Search for Current Information
How to use DeepSeek's Search switch: when to turn it on for news, prices and recent facts, how to ask for sources, and why you still need to check.
- 11
How to Sign Up for DeepSeek and Fix Login Problems
How to sign up for DeepSeek and fix login problems: missing codes, rejected email domains, Google or WeChat errors, and suspended accounts.
- 12
How to Use DeepSeek Safely at Work: A Checklist for Businesses
How to decide whether your business can use DeepSeek: data stored in China, government restrictions, and safer setups like local models and clear rules.
- 13
How to Solve Math Problems With DeepSeek (and Check the Answer)
How to use DeepSeek for math: turn on DeepThink, give the full problem, use the step-by-step prompt DeepSeek recommends, and check the answer.
- 14
How to Run DeepSeek in LM Studio Without the Command Line
How to run DeepSeek-R1 in LM Studio: check the requirements, download a model from 4.3 GB, chat offline, and chat with your own documents privately.
- 15
How to Write Better DeepSeek Prompts for Reasoning and Search
How to write prompts that suit DeepSeek: the whole problem up front for DeepThink, sources with Search, no system prompt on local R1, and a clear format.
- 16
How to Fact-Check DeepSeek's Answers Before You Use Them
How to fact-check DeepSeek: what its own terms say about errors, using Search for sources, checking quotes, and labeling AI content when you publish it.
- 17
How to Use DeepSeek in Claude Code, OpenCode and Other Apps
How to use DeepSeek in other apps: connect Claude Code, OpenCode or any OpenAI-compatible tool with an API key, or use a local model through Ollama.
- 18
How to Read the DeepSeek Open-Weight License Before You Build
How to check a DeepSeek model's license before you build: MIT for recent models, separate base-model licenses for distilled R1 versions, and where to look.
- 19
How to Share, Export and Delete Your DeepSeek Chats
How to manage DeepSeek chats: share one as a link or image, export all your data, rename and pin chats, and delete chats or your whole account.
- 20
How to Cut Your DeepSeek API Costs: Off-Peak Hours and Caching
How to spend less on the DeepSeek API: run jobs off-peak at half price, reuse the same prompt opening for cache hits, and switch off thinking when not needed.
Using AI to write code? See the AI coding guides and the AI glossary.