deepseek-harness-desktop
为 DeepSeek Harness (DSH) 生态打造的现代化桌面端体验
The Lens
By Erik Loyd, SaaS CEO and former COO/CFO of an AWS Premier Partner.
Updated Aug 2026
DSH Desktop turns DeepSeek Harness into a normal app. The harness is DeepSeek's open source agent runtime, and running it means Node, pnpm, a terminal, and a port. This is one installer for Windows x64 or an Apple Silicon Mac, and the Web UI, host service, and plugin system open in a native window. MIT, free, and not an official DeepSeek project.
Nothing to host. The installer bundles Electron, Node, pnpm, and a pinned harness build, and leaves your global PATH alone. Updates download only after you confirm. Upstream runs unmodified, so official plugins install from the bundled terminal.
Solo and small teams on Windows or an M-series Mac should use this instead of the npx command. Medium and large teams get nothing extra: no shared profiles, no policy controls. Anyone already running the harness in a browser gets more from zhu1090093659/dsh-web-ui. For a multi-provider chat client instead, CherryHQ/cherry-studio is further along.
The catch is what it pins. Upstream calls itself a developer preview and warns in capitals that breaking changes are coming, and this ships a fixed release candidate. No Linux installer, no Intel Mac, and the issue tracker carries launch crashes on both supported platforms.
Free vs Self-Hosted vs Paid
fully freeFree
Everything. MIT licensed, no accounts, no paid edition. The v2.0.0 installers are about 148 MB for Windows x64 and about 334 MB for Apple Silicon macOS, published on GitHub Releases with SHA-256 digests and mirrored at dshdesktop.cn. The maintainers state the project is free and that anyone trying to sell it should be refused.
Paid
There is no paid tier for the app itself. The bill comes from whichever model you point it at. DeepSeek switched to peak and off-peak API billing on August 16, 2026: V4-Pro runs roughly $0.66 per million input tokens and $1.98 per million output off-peak, about double that in peak windows. V4-Flash runs roughly $0.22 in and $0.66 out off-peak. Both were flat before the change, at $0.435 / $0.87 and $0.14 / $0.28.
Self-Hosted Costs
The app is the self-hosted path, so there is no server bill. Disk is the real line item, since the installer ships Electron, Node, pnpm, and a pinned harness. The harness also accepts Anthropic, OpenAI, Bedrock, Vertex, Azure, and any custom OpenAI-compatible endpoint, so pointing it at a local vLLM or Ollama server drops token cost to electricity.
When to Pay
Never for the app. Pay for tokens, and schedule long agent runs off-peak now that DeepSeek bills by the clock. V4-Flash off-peak is the sane default for work that is cheap to verify.
Free forever with no paid edition. The only bill is model tokens, and since DeepSeek went peak and off-peak on August 16, 2026, when you run changes what you pay.
What to do by team size
- Solo
- free, and it beats running npx and babysitting a local port
- Small team
- free; everyone installs their own copy, since profiles are per machine
- Medium team
- free, but expect no shared config, no policy controls, no seat management
- Large team
- free, though nothing here helps you govern model access or token spend
Get tools like this every Wednesday
One featured tool, three on the radar. No fluff.
Similar Tools
an open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM
The open source coding agent.
⌥ AI Coding agent for the terminal — hash-anchored edits, optimized tool harness, LSP, Python, browser, subagents, and more
An open-source AI coding agent that lives in your terminal.
Autonomous coding agent right in your IDE, capable of creating/editing files, executing commands, using the browser, and more with your permission every step of the way.
DeepSeek + MiMo coding agent in terminal
A low score is not a verdict on quality. Young and niche tools start low by design. How we calculate scores
Trust Signals
License: MIT License
Use freely, including commercial. Just keep the license.
Commercial use: ✓ Yes
About
- Owner
- anywhere-labs (Organization)
- Stars
- 9,572
- Forks
- 403
Also by anywhere-labs
Explore Further
More tools in the directory
lobehub
🤯 LobeHub is your Chief Agent Operator, organizing your agents into 7×24 operations by hiring, scheduling, and reporting on your entire AI team.
81.8k ★unsloth
Unsloth is a local UI for training and running Kimi K3, Gemma 4, Qwen3.6, DeepSeek-V4, GLM and other models.
72.7k ★cherry-studio
AI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs
50.6k ★