Flyvhidbwo/dsh-vision-proxy

dsh-vision-proxy

English | [绠€浣撲腑鏂嘳(README.md)

Keep DeepSeek as the brain 鈥?paste images anyway. GUI image attachments auto-transcribed for text-only DeepSeek on DeepSeek Harness.

⚠️ Compatibility & positioning (Aug 2026): this plugin now supports dsh 0.1.1-rc.2 (adapter prepareCall contract). Since dsh 0.1.1, DeepSeek natively supports multimodal models (e.g. DeepSeek-V4-Flash-Vision-Exp) — if you use an official vision model, just attach images, no plugin needed. DeepSeek-V4-Pro / plain Flash are still text-only — use this plugin for image transcription with them (only Flash-Vision-Exp is natively multimodal). The plugin also serves: local Ollama (images never leave the machine, free), custom OpenAI-compatible VLM setups.

Why this exists

DeepSeek Harness natively gates image attachments on the selected model's declared inputModalities. DeepSeek's chat-completions line is text-only, so attaching an image with DeepSeek selected is rejected by design. Tool-based vision plugins exist, but GUI image attachments still fail with a text-only model.

This plugin closes that gap: it registers a new provider route (deepseek-vision) that wraps the real DeepSeek adapter, claims image input (so the preflight admits attachments), and transcribes every attached image to text in the request stream before delegating to DeepSeek. The conversation is still answered by DeepSeek; vision is an add-on.

user attaches image 鈹€鈹€鈻?deepseek-vision route 鈹€鈹€鈻?transcribe via VLM (OCR + layout + details)
                          鈹?                         鈹?
                          鈻?                         鈻?
                   DeepSeek answers 鈼€鈹€鈹€ text-only conversation (images replaced by [鍥剧墖杞瘧] text)

Features

• No hangs, ever. Anonymous endpoints are hard-capped at 20 s (a hanging free tier can no longer stall a turn for minutes); HTTP 429 on anonymous endpoints fails immediately instead of sleeping on a useless Retry-After; endpoints that just failed (429/timeout) are cooled down for 60 s and skipped.
• Multi-model, multi-provider. Any OpenAI-compatible VLM endpoint works 鈥?DashScope/Qwen, QwenCloud (international), Zhipu, OpenRouter, local Ollama, or your own. Each fallbackModels entry can carry its own baseURL/model, so one install can chain providers.
• Zero-config local path. With autoLocalOllama (default on), a running Ollama at http://localhost:11434 is detected at startup and prepended to the fallback chain 鈥?images never leave your machine. No key, no account.
• Fast, clear failures. With no key and no local Ollama, transcription fails in seconds with actionable guidance (configure VISION_API_KEY / DASHSCOPE_API_KEY or install Ollama) 鈥?never a silent stall.
• Automatic upgrade when you have a key. Export VISION_API_KEY / DASHSCOPE_API_KEY and your configured paid endpoint is used automatically (default: DashScope qwen3.7-flash 鈥?fast, cheap, no rate limit; DashScope, QwenCloud, Zhipu, OpenRouter, or any OpenAI-compatible endpoint all work); keyless entries are skipped, not failed.
• Install-time consent prompt. postinstall asks whether you have a VLM API key. Non-interactive environments skip the prompt; the install never hangs. A PRIVACY NOTICE is printed at startup naming the active endpoint.
• Fallback chain with classified errors. rate_limit / quota / auth / region / model_not_found / context_too_large / http are classified with actionable hints.
• Content-hash cache. Transcriptions are cached by the SHA-256 of the image bytes (in-process, capped at 200) 鈥?the same image is transcribed at most once per process, even re-attached or in another conversation.
• Auto-downscale (optional). With sharp installed, images above maxImagePixels are downscaled before transcription 鈥?fewer image tokens, much faster on big screenshots. Degrades gracefully without sharp.
• read_image compatible. The native read_image tool also works on this route (its capability gate reads the same model info).

Supported models & providers

One config (baseURL + model, optionally apiKey) covers every backend:

Scenario	baseURL	model	Notes
DashScope (China) 鈥?default main	https://dashscope.aliyuncs.com/compatible-mode/v1	qwen3.7-flash / qwen3-vl-flash	Cheap, fast, no rate limit. Keys: sk-ws-鈥 from [platform.qianwenai.com](https://platform.qianwenai.com) or sk-鈥 from bailian.console.aliyun.com
Local Ollama (auto-detected)	http://localhost:11434/v1	first vision-capable model	Zero config when installed; images never leave the machine
QwenCloud (intl.)	https://dashscope-intl.aliyuncs.com/compatible-mode/v1	qwen3-vl-plus etc.	International variant
Zhipu (free tier)	https://open.bigmodel.cn/api/paas/v4	glm-4.6v-flash	Free tier, still needs a (free) Zhipu API key
Anything OpenAI-compatible	your endpoint	your model	OpenRouter, Ark, vLLM, gateways鈥?the plugin only speaks /chat/completions

鈿狅笍 Anonymous third-party free tiers are NOT bundled as a default fallback. In field testing, anonymous free endpoints (e.g. OVHcloud AI Endpoints) were strictly rate-limited AND occasionally hung without a response 鈥?as a default they just reproduce a broken experience. If you still want to point at one, add it yourself via fallbackModels with anonymous: true (the 20 s cap still applies).

Pricing (CNY, Alibaba Model Studio mainland + DeepSeek official, Aug 2026 reference)

Model	Input	Output	One 1080p screenshot (≈2000 tokens)
qwen3-vl-flash	¥0.15 / 1M tokens	¥1.5 / 1M tokens	≈ ¥0.0005 (0.05 fen)
qwen3.7-flash	¥0.2 / 1M tokens	¥0.8 / 1M tokens	≈ ¥0.001 (0.1 fen)
deepseek-v4-flash-vision-exp (official, default)	¥3 / 1M (peak) · ¥1.5 (off-peak)	¥9 / 1M (peak) · ¥4.5 (off-peak)	≈ ¥0.01 (peak) / ¥0.005 (off-peak)
Local Ollama	free	free	¥0 (images never leave the machine)

Transcription prefers the official deepseek-v4-flash-vision-exp by default (peak hours 9:00-12:00 / 14:00-18:00 CN at full price, half-price otherwise; images convert to tokens by size, no per-image fee; about 1/3 of the V4-Pro tier price). Images are billed by token (a 1080p screenshot ≈ 2000 tokens). Even heavy use (100 images/day) is only a few yuan a month. Local Ollama is completely free. Always check provider consoles for live pricing.

Key resolution order: config apiKey 鈫?$VISION_API_KEY 鈫?$DASHSCOPE_API_KEY. Anonymous endpoints (anonymous: true) and local hosts need no key; keyless non-anonymous entries are skipped automatically.

Quick start

dsh plugin --profile web add dsh-vision-proxy

During install you are asked one question 鈥?do you have a VLM API key? Answer y for the paid fast path, or N (default) for the local/zero-config path. Restart dsh web, pick DeepSeek + 鑷姩璇嗗浘 in the model selector, then paste an image into any conversation.

pnpm 鈮?10 blocks dependency build scripts by default 鈥?the first install exits non-zero with Ignored build scripts: dsh-vision-proxy, sharp. Approve both (the plugin's consent prompt and sharp's optional binary), then re-run the install to finish bundle registration:

# in the profile's pnpm-workspace.yaml
allowBuilds:
  dsh-vision-proxy: true
  sharp: true

dsh plugin --profile web add dsh-vision-proxy   # re-run after approving

Slow npm registry in China? dsh plugin --profile web add dsh-vision-proxy --registry=https://registry.npmmirror.com (the flag is forwarded to pnpm).

Live demo: a real GUI image turn

A real conversation on the deepseek-vision route (DeepSeek-V4-Flash as the brain): the user pasted a meme and asked "浣犵湅鍒颁簡浠€涔? (what do you see?); the image was auto-transcribed by the VLM and DeepSeek answered from the text 鈥?one step, ~7.6 s.

Left: the model picker showing the deepseek-vision route (DeepSeek + 鑷姩璇嗗浘) selected 鈥?that is what admits image attachments. Right: DeepSeek's full answer derived from the transcribed image text.

user pastes a meme image + "浣犵湅鍒颁簡浠€涔?
  鈫?image block auto-transcribed via the VLM (OCR + layout):
      "鎴戞槸鍚冪櫧楗殑 / 钃濊壊澶ц偉楸硷紒 (鐞嗙洿姘斿．.jpg) 鈥?Q-version blue-haired maid girl
       with a whale tail, holding a bowl of rice and chopsticks, excited expression"
  鈫?DeepSeek answers with a full visual analysis of the meme

Two autonomous paths are covered: the view_image tool (any route, file paths & URLs) and image-block auto-transcription (on the deepseek-vision route 鈥?images you attach mid-conversation).

Configuration

The bundle already ships sensible defaults (see the strategy above) 鈥?you normally don't need to configure anything. To override them in your profile, use an id-targeted override, NOT an insert (see the warning below):

# $DSH_HOME/profiles/web/cordis.patch.yml 鈥?user-layer override example
- id: dsh-vision-proxy
  name: 'dsh-vision-proxy'
  config:
    baseURL: https://dashscope.aliyuncs.com/compatible-mode/v1
    apiKey: 'sk-鈥?          # or leave '' to read env vars (writing it here is the reliable way on Windows)
    model: qwen3.7-flash
    maxTokens: 4096
    timeoutMs: 120000       # anonymous endpoints are hard-capped at 20 s anyway
    maxImagePixels: 4000000
    marker: '[鍥剧墖杞瘧]'
    autoLocalOllama: true
    fallbackModels: []      # add your own {model, baseURL, apiKey?, anonymous?, timeoutMs?}

鈿狅笍 Do NOT write this as - insert: [{id: dsh-vision-proxy, 鈥]. In dsh's patch semantics an insert appends entries to the list 鈥?the bundle's own entry and yours (same id) would both be instantiated, registering the deepseek-vision adapter twice (undefined behavior). A top-level - id: entry targets the existing row and replaces its whole config; keys you omit fall back to the plugin schema's .default() values (e.g. maxTokens=4096, timeoutMs=120000, autoLocalOllama=true), so writing only apiKey/model also works.

Key	Default	Meaning
providerId	deepseek-vision	Route id shown in the model picker
innerProvider	deepseek-official	Existing adapter route to wrap
baseURL	DashScope compatible-mode	OpenAI-compatible VLM endpoint (any vendor, Ollama included)
apiKey	''	VLM key; falls back to $VISION_API_KEY, then $DASHSCOPE_API_KEY. On Windows, environment changes may not reach a running dsh 鈥?writing apiKey here is the reliable way
anonymous	false	Skip the Authorization header (for registration-free endpoints; 20 s timeout cap applies)
model	qwen3.7-flash	Vision model id (e.g. Qwen2.5-VL-72B-Instruct, qwen3-vl-flash, glm-4.6v-flash, qwen3-vl:4b)
maxTokens	4096	VLM output cap (thinking models spend tokens on reasoning first)
timeoutMs	120000	VLM request timeout (anonymous endpoints are capped at 20 s regardless)
maxImagePixels	4000000	Images above this are downscaled before transcription when sharp is installed (0 disables)
marker	[鍥剧墖杞瘧]	Marker prepended to each transcription
failureMode	placeholder	Behavior when every VLM fails for an image: placeholder (default) inserts [图片转译失败: ...] text and the conversation continues - a dead endpoint can no longer poison the session; error fails the whole turn (legacy)
autoLocalOllama	true	Probe http://localhost:11434 at startup; when found, prepend it to the fallback chain
localOllamaModel	''	Ollama model id; empty picks the first vision-capable model the local Ollama reports
fallbackModels	[]	Ordered fallback list {model, baseURL?, apiKey?, anonymous?, timeoutMs?} 鈥?each entry may point at a different provider; keyless non-anonymous entries are skipped

About API keys on Windows: dsh --profile <name> --dump-config prints the composed config as-is (so a key in cordis.patch.yml shows in plaintext dumps), but environment variables set after a process started (explorer.exe caches them) may never reach a running dsh. If you see skipped 鈥?no API key despite having exported the key, write apiKey directly into the plugin config 鈥?it is the only reliable path on Windows. (Note: dsh rc.6 does NOT load .env files, so that is not an alternative.)

Verify the install

dsh --profile web --dump-config | grep -A3 dsh-vision-proxy   # exactly ONE entry (note: dumps config in plaintext, key included)

1. Restart dsh web 鈫?the model picker shows DeepSeek + 鑷姩璇嗗浘.
2. Paste an image into a conversation 鈫?you should see the [鍥剧墖杞瘧] marker followed by DeepSeek's answer.
3. With no key and no local Ollama, the turn should fail fast (seconds) with the guidance message 鈥?that is the intended no-hang behavior.

Behavior notes

• Only messages containing image blocks are touched; plain-text conversations hit DeepSeek with zero overhead.
• Anonymous endpoints: 20 s hard timeout cap, HTTP 429 fails immediately (no retry), failures arm a 60 s endpoint cooldown 鈥?consecutive images don't re-hit a broken endpoint.
• The request only fails after every chain entry failed, with one error listing each attempt plus actionable guidance.
• Transcription results are cached in-process by image content hash (never persisted).
• On startup the plugin logs a one-line summary 鈥?route id, wrapped provider, VLM model, endpoint, timeout, maxTokens, apiKey source and fallback list (the key itself is never logged), plus a PRIVACY NOTICE and a local-Ollama detection line.
• Tested: 14 unit tests on Node 22 and 24 via GitHub Actions (incl. no-hang fast-fail, cooldown skip, and Ollama detection).
• Transcription quality: dense UI screenshots may lose small text details 鈥?that is the vision model's capability ceiling, not a plugin bug. For OCR-heavy work, use a stronger model (e.g. qwen3-vl-plus) or raise maxTokens.

Troubleshooting

Symptom	Cause & fix
skipped 鈥?no API key despite exporting VISION_API_KEY	Windows caches environment variables in explorer.exe; the running dsh never saw them. Write apiKey directly into the plugin config, then restart dsh
Ignored build scripts: dsh-vision-proxy, sharp on install	pnpm 鈮?10 blocks dependency build scripts. Add allowBuilds: {dsh-vision-proxy: true, sharp: true} to the profile's pnpm-workspace.yaml, then re-run the install
ERR_PNPM_MINIMUM_RELEASE_AGE_VIOLATION on a fresh release day	pnpm 11 defaults minimumReleaseAge to 1 day (supply-chain policy). Add minimumReleaseAge: 0 to the profile's pnpm-workspace.yaml, or pass --config.minimum-release-age=0 to dsh plugin add, then re-run
all N vision model(s) failed 鈥?rate_limit on an anonymous endpoint	Anonymous free tiers are strictly rate-limited and may hang. Configure a key or use local Ollama
Turn stalls ~20 s then fails on a fresh install with no key	No key and no local Ollama 鈥?that is the intended fast-fail path. Install Ollama or add a key
Slow downloads from registry.npmjs.org	Use --registry=https://registry.npmmirror.com (forwarded to pnpm)

Privacy

Transcription sends image bytes (base64, over HTTPS) to the configured VLM endpoint 鈥?the image data leaves your machine unless baseURL points at a local service (e.g. Ollama). Nothing is stored beyond the harness's own attachment store. For sensitive images, use your own endpoint or a local model 鈥?or don't install.

How it works (for plugin developers)

The plugin uses only public harness seams, stable on rc.6:

• ctx.llm.registration(innerProvider).adapter 鈥?reach the wrapped adapter;
• ctx.llm.registerAdapter([providerId], proxyAdapter) 鈥?register a NEW route (no DUPLICATE_ADAPTER conflict);
• proxy resolveModel overrides inputModalities to ['text', 'image'] 鈥?satisfies the attachment preflight (api-proxy) and the read_image gate (dsh-tool-fs);
• proxy stream transcribes image blocks (shape { type: 'image', attachment }, bytes via ctx.get('attachments').readImage(ref)) and yield*s the inner adapter's stream unchanged.

License

MIT