huashenglian/dsh-her-eyes
41Last commit Aug 13, 2026
dsh-her-eyes DSH plugin
dsh-her-eyes is a dual-face plugin for DeepSeek Harness. It registers an analyze_image tool backed by primary/backup OpenAI-compatible vision endpoints with automatic failover. It also adds an auto-saving Settings → Vision Models (VLM) page that follows the harness UI language.
How to install the dsh-her-eyes DSH plugin
dsh plugin --profile web add ./dsh-her-eyesCopying does not run this command. Review the repository and version before installing the dsh-her-eyes DSH plugin.
dsh-her-eyes DSH plugin data source
dsh-her-eyes DSH plugin snapshot date: Aug 16, 2026
discovered
What the dsh-her-eyes DSH plugin can do
- Provides an analyze_image tool available in every session via the global tools registry.
- Supports primary and backup OpenAI-compatible VLM endpoints with automatic failover after consecutive failures.
- Adds an auto-saving settings page for Vision Models (VLM) under Settings, with model list fetching from the endpoint.
- Registers web routes /vlm/config, /vlm/models, /vlm/reset served by the host half.
- Supports i18n: the settings page language matches the harness UI (English or Chinese).
Where the dsh-her-eyes DSH plugin fits
- Analyze images in conversations using a vision-language model through the analyze_image tool.
- Configure multiple VLM backends (primary and backup) for reliability and automatic failover.
- Manage VLM settings directly from the harness UI without manual config file editing.
- Fetch available model list from the VLM endpoint via the settings page.
Who the dsh-her-eyes DSH plugin is for
- Developers using DeepSeek Harness who want to integrate image analysis capabilities.
- Users who need a reliable VLM tool with failover support for AI sessions.
dsh-her-eyes DSH plugin limitations
- Requires a dsh CLI with a web profile installed and pnpm on PATH (or npx with pnpm).
- Requires self-provided OpenAI-compatible API endpoints and API keys for VLM services.
- Only supports vision models that are compatible with OpenAI's chat completions API format.
- The settings page auto-saves immediately; no manual save button is available.
dsh-her-eyes DSH plugin: from the repository README
Quoted from the huashenglian/dsh-her-eyes README, the upstream source of the dsh-her-eyes DSH plugin. Copyright remains with the original authors.
English | [中文](README.zh.md) A Vision-Language-Model (VLM) analyzer plugin for **DeepSeek Harness** (`dsh`). It gives the AI an `analyze_image` tool backed by primary/backup OpenAI-compatible vision endpoints with automatic failover, and adds an auto-saving **Settings → Vision Models (VLM)** page (English/中文). ## Features - **`analyze_image` tool** registered on the global tools registry — available in every session. - **Primary / backup VLM APIs** — OpenAI-compatible `endpoint` + `apiKey` + `model`; if the primary fails consecutively beyond the retry count, the AI automatically falls back to the backup. - **Auto-saving settings page** — every edit saves and takes effect immediately; no Save button. Also lets you fetch the model list from the endpoint. - **Web routes** `/vlm/config`, `/vlm/models`, `/vlm/reset` served by the host half. - **i18n** — the settings page follows the harness UI language (English / 中文). ## Requirements - `dsh` CLI (DeepSeek Harness) with a `web` profile installed. Requires `pnpm` on `PATH` (or use `npx --yes pnpm@<version>`). ## Install The plugin is a **bundle**: it carries its own `cordis.patch.yml` and self-activates. Installing it is a single c
Read the full READMERepository license: MIT
dsh-her-eyes DSH plugin questions
How do I install dsh-her-eyes?
You can install it using the dsh CLI: run `dsh plugin --profile web add github:huashenglian/dsh-her-eyes`. Make sure you have a web profile set up and pnpm available on PATH. If pnpm is not on PATH, you can use `npx --yes pnpm@11.7.0` instead.
What API keys do I need to configure?
You need an OpenAI-compatible API key for a vision-language model. Configure the primary and optionally a backup endpoint in the settings page or in the `vlm-vision.json` file. The API key is masked when saved via the settings page.
Does the plugin support multiple vision models?
Yes, you can set any model that is compatible with your OpenAI-compatible endpoint. The plugin also lets you fetch the model list from the endpoint via the settings page. You can switch between primary and backup models automatically if the primary fails.
How does the automatic failover work?
The plugin tracks consecutive failures. If the currently active API (primary or backup) fails `retryCount` times in a row, it switches to the other API. You can configure `retryCount` in the settings or in `vlm-vision.json`.
Can I use dsh-her-eyes without a web profile?
No, the plugin requires a web profile because it registers web routes and a settings page that depend on the harness web server. The install command `dsh plugin --profile web add` already assumes a web profile.