liustack / modlens
git clone https://github.com/liustack/modlens.git
v3.27.1 · 18d ago
Vision plugin that gives text-only coding agents and DeepSeek Harness models image understanding by converting pasted images into structured JSON evidence.
Why you might care
Lets text-only models like DeepSeek and GLM read images pasted into chat, returning OCR, layout, and semantic data as JSON. Installs as a DeepSeek Harness plugin with one command and auto-wraps eligible text-only routes with a vision-enabled model entry.
+19
stars, 24h · #2,226
+217
stars, 7 days · #1,411
+~784
stars, 30 days · #1,373
4.7k
★ total · +5% this week
148
forks · 3 open issues
Star history (as recorded here)
LanguageTypeScript
LicenseMIT
Created2026-02-22 (8mo ago)
Last push10d ago
Watchers5
OwnerUser
GitHub topics
agent-skillsclaude-codeclaude-skillscodexcordisdeepseekdshdsh-pluginglmharnessharness-engineeringhermes-agentimage-to-textmultimodalocropenclawpi-agenttext-only-llmvisionvision-transformer
Links
- GitHub: github.com/liustack/modlens
- Homepage: https://liustack.dev/tools/modlens
- Releases · Issues · Hacker News mentions
README
Loading README…
Blurb generated by claude-haiku-5-5 6d ago. Star deltas marked ~ are extrapolated from partial history.