renezander030

Browserground — AI skill for Claude Code

AI community

Local UI-grounding specialist for hybrid AI agents.

How to install Browserground

This entry records only its repository, not the path inside it, so there is no exact command to give. Open renezander030/browserground and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Browserground does

Local UI-grounding specialist for hybrid AI agents. Qwen3-VL-2B LoRA. Screenshot + text target → strict JSON bbox. Drop-in for Claude Code, Codex, browser-use. Cuts GPT-4V cost & latency.

Alternatives in AI

  • Qwen Code — A command-line AI workflow tool adapted from Gemini CLI, optimized for Qwen3-Coder models with enhanced parser 20.8k ★
  • Anthropic Quickstarts — by Anthropic - Offers comprehensive development guides for three distinct AI-powered demo projects with standa 15.4k ★
  • Surface — Show ranked attack surface for a target from its recon manifest + hunt memory 3.8k ★

README

browserground v0.3 — local UI-grounding specialist for hybrid AI agents. MLX 4-bit, npm, pip, Ollama. ScreenSpot-v2 60%. Strict JSON output.

browserground

The local UI-grounding specialist for hybrid AI agents.
Drop in a screenshot + text target, get a strict JSON bbox. 2B params. MLX-native. Apache 2.0.

HF model MLX build GGUF build npm PyPI License

TL;DR — when to use browserground (and when to use UI-TARS-MLX instead)

If you're on Apple Silicon with ≥16 GB RAM and you need **generic, max-accuracy UI grounding**, use **[mlx-community/UI-TARS-1.5-7B-4bit](https://huggingface.co/mlx-community/UI-TARS-1.5-7B-4bit)**. It's the obvious default — ~94% on ScreenSpot-v2, MLX-native, drops into `mlx-vlm` directly. ByteDance research-lab compute, you couldn't reproduce it on a budget.

browserground is for two narrower jobs:

1. The recipe for *your product's* custom UI grounder

UI-TARS is a finished model. You can use it; you can't easily extend it. The training

Related Skills

LLM Evaluation Framework — AI skill for Claude Code

LLM Evaluation Framework

Production-grade LLM Evaluation & Benchmarking Framework - GPT-4, Claude, Gemini, Mistral. Accuracy, latency,

AI community
Gemini Model Router — AI skill for Claude Code

Gemini Model Router

Local-first agentic terminal that routes prompts across local Gemma 4 (vLLM), Gemini CLI, and Claude Code — pi

AI community
Code Context Engine — AI skill for Claude Code

Code Context Engine

Local-first MCP context engine for AI coding agents — cuts token cost ~90% via a per-project code graph + code

AI community
Semble Rs — AI skill for Claude Code

Semble Rs

Fast, AI-agent-native code search in Rust — hybrid BM25 + semantic, Tree-sitter AST chunking, dependency & imp

AI community
Hipocampus — AI skill for Claude Code

Hipocampus

Drop-in memory harness for AI agents — 3-tier memory, compaction tree, hybrid search. One command to set up. W

AI community
Doc Search MCP — AI skill for Claude Code

Doc Search MCP

Local MCP server that indexes PDFs, EPUBs, HTML, Markdown and text files and makes them searchable by LLM codi

AI community

Related Agents

Gemini GPT Hybrid Hard — Development agent for Claude Code

Gemini GPT Hybrid Hard

AGGRESSIVE hybrid agent that delegates code generation and modification directly to Gemini and GPT for rapid d

Development community
Uxaudit L3 Judge — Development agent for Claude Code

Uxaudit L3 Judge

L3-vision Judge for the uxaudit pipeline. Reads ONE captured screenshot plus a check-specific prompt.md (and o

Development community
Brain Eval Engineer — Testing & QA agent for Claude Code

Brain Eval Engineer

Evaluation engineer — question set tooling, layered metrics (harvest/graph/retrieval/answer), fixed-strategy a

Testing & QA community