Step-by-Step Guide to Installing DeepSeek Harness (dsh)
This tutorial walks you through installing DeepSeek Harness (dsh), covering Node.js version checks, quick one‑command setup, source‑code installation with pnpm, Docker options, and initial API key configuration for the web UI on your local machine.
DeepSeek Harness (dsh) Overview
DeepSeek Harness (dsh) is an open‑source agent framework that enables large language models to operate on local files, execute commands, and invoke tools. It is in developer preview and requires Node.js ≥ 22.
Node.js Environment Check
node --versionVersion output v22.x.x or higher → ready.
If the command is missing or the version is lower, install the latest LTS from the Node.js website, reopen the terminal, and re‑run.
Quick One‑Command Experience
npx @deepseek-ai/dsh webFirst execution downloads dependencies (may take 1–2 minutes). When the terminal prints an address such as http://127.0.0.1:3080, open it in a browser.
⚠️ If the process hangs or is very slow, the cause is typically network latency to the npm registry. Switch to a domestic npm mirror or use the source‑code installation method.
Source Installation (Customization / Development)
Clone the repository
git clone https://github.com/deepseek-ai/deepseek-harness.git
cd deepseek-harnessInstall dependencies and build
pnpm install
pnpm run buildStart the service pnpm dsh web Open http://127.0.0.1:3080 in a browser.
Alternative Options
Community Docker image
docker run -d --name dsh -p 3080:3080 ghcr.io/huoxue1/deepseek-harness:latestCommunity desktop version – a pre‑packaged Windows binary available on GitHub; unzip and run.
First‑Run Configuration – API Key
Open Settings in the left sidebar.
In the Model section, click “Add Provider” or “Add Custom Provider”.
Enter the API key. Options:
Official DeepSeek Open Platform – create and copy the key.
Other compatible platforms (e.g., Baidu Qianfan) – select “Custom Provider” and provide the endpoint and key.
Save the settings, return to the main screen, select a workspace directory (the project folder the AI should operate on), choose a model, and start chatting.
Tips
Run modes – preset modes such as “Standard” (full features) and “Minimal” (suitable for benchmarking) can be switched in the UI.
Network issues – when installing from source, configure npm or pnpm to use a domestic mirror if dependency downloads are slow.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
Coder Trainee
Experienced in Java and Python, we share and learn together. For submissions or collaborations, DM us.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
