chat-deepseek-vision

Configuration

Plugin config comes from the loader entry's config; unset fields fall back to the same-named environment variables:

Plugin config	Env var	Default	Description
token	DSW_TOKEN	-	DeepSeek Web login token (64 chars, see below)
cookie	DSW_COOKIE	-	Session cookie (e.g. ds_session_id=..., optional)
baseUrl	DSW_BASE_URL	https://chat.deepseek.com	DeepSeek Web endpoint
timeoutMs	VISION_TIMEOUT_MS	180000	SSE idle timeout (ms): measured as "no new data" duration, so long outputs are not killed by a total-time cap
screenshotDir	VISION_SCREENSHOT_DIR	~/Pictures/Screenshots	Screenshot save directory
autoDescribe	-	true	Auto-describe send-box attachments and inject the description; false keeps only the analyze_attachment manual tool
attachmentPrompt	-	Default describe prompt	Custom prompt for auto-describing attachments
nativeImageModels	-	false	When true, if the current model declares image input (e.g. pi-ai vision models) the plugin skips projection/description and uses the native image pipeline
userAgent	-	Chrome 132 UA	Request UA (optional)

Getting the token (64 chars):

1. Log in to chat.deepseek.com in a browser
2. F12 → Console, run: JSON.parse(localStorage.getItem('userToken')).value
3. Fill the output into token (or DSW_TOKEN)

Limits: images only (screenshots, local images, URLs, data URIs), single image ≤ 15MB.