DeepSeek Harness (dsh)
DeepSeek's official open-source agent harness with a local Web UI.
DeepSeek Harness runs a local Web UI and connects to DeepSeek, built-in third-party providers, or a custom model API. Launch it from your project directory, configure a model, then select that directory as the workspace.
Setup steps
- 1
Launch the Web UI
Requires Node.js 22.19 or newer within the 22.x line, or Node.js 24 or newer. npx downloads the package and starts the local Web UI at http://127.0.0.1:3080; no global installation is required.
bashnpx @deepseek-ai/dsh web - 2
Configure DeepSeek
Open Settings → Models, enter a DeepSeek API key, and save it. The model becomes available for the next request without restarting dsh.
- 3
Optional: add another model API
Choose Add model provider. Select a built-in provider, or create a custom API with a lowercase provider ID, base URL, protocol, API key, and at least one model ID.
- 4
Choose the workspace
Choose workspace, add the project directory where you started dsh, and select it. The session composer remains unavailable until a workspace is selected.
- 5
Start a session
Select the provider and model, then create a session. Model configuration changes apply to the next request; a session that has already sent a request retains its recorded model.