54xkeee/dsh-vision

README excerpt

👁️ dsh-vision Eyes for text-only DeepSeek on DeepSeek Harness — powered by Doubao Web by default, zero cost, no API key. 简体中文 =20" / TL;DR : no API key, no paid tier — log into Doubao once in your browser, and DeepSeek can see images in every conversation. Paste, recognize, answer. ✨ Why you'll love it Pain point dsh-vision solution Vision APIs cost money and need keys Doubao Web by default — zero cost, no API key, just a browser login DeepSeek is text-only; pasting an image is rejected Wrapper adapters claim image input; images become text placeholders automatically Other plugins lock you into one vendor Doubao Web (default) + Antigravity IDE quota (flash/pro) + Gemini API + Cockpit proxy — auto fallback chain The model "forgets" what it saw Vision evidence memory : results persist in the session, reused across turns, restored after compaction Paying to re-recognize the same image Content-hash cache : same image + same question = recognized once per process Complex images get shallow answers Auto detail escalation : standard pass first, auto-upgrade to deep for complex scenes WSL / firewalled networks winCurl fallback for API channels; Doubao Web channel runs through your Windows…

Category