amarnath3003/quickliquid
Liquid Glass - Heavily Optimised.
GITHUB SIGNALS / PROJECT INDEX
Search the archive. Follow your curiosity. Keep the projects you want to come back to.
Liquid Glass - Heavily Optimised.
The capability router for AI coding agents. Route any task to a bounded skill portfolio across Claude Code, Codex, Cursor, Jcode, Hermes and more — with a built-in prompt-injection firewall.
Convert and Run Exl3 format LLM quant on Apple Silicon.
A visual, non-linear conversation workspace plugin for DeepSeek Harness ; A canvas-based session explorer and branching workspace for DeepSeek Harness.
A SDK for building MCP tools, prompts, resources, and widgets.
Create Datasets , FineTune Models , Test Models - All in one place. Everything simplified
OpenClaw but for hardware and IoT devices.
Windows maintenance CLI + TUI that sifts the junk from the keep. Dry-run by default, Recycle Bin only, undo included.
Day-0 deployment of Qwen3.8-Flash-Next (NVFP4) on 2x DGX Spark TP2 via SGLang. Includes the SM121 QSA kernel-guard fix, launcher, benchmarks, and full report.
High-performance Cloudflare Workers–compatible infrastructure in a single binary
Learn PyTorch by building your own. (inspired from Harvard's TinyTorch)
Local-first WeChat intelligence system with a read-only CLI, Codex skills, searchable chat history, daily briefings, follow-ups and opportunity tracking.
Use ChatGPT (Codex), Claude, and Grok (X Premium) subscriptions as DeepSeek Harness LLM providers — OAuth login in the web UI, no API keys
Splits text into lines, words and characters exactly where the browser already broke it.
Open the project to explore the idea and the code.
Native local Taskboard plugin for DeepSeek Harness. SQLite-backed projects, Agent claim/review, and a native Web UI — no iframe, no second chat runtime.
HRM-Text is a 1B text generation model based on the HRM architecture, strengthened by task completion and latent space reasoning.
Personal Optimized SGLang runtime for Qwen3.8-27B/DFlash2 and Qwen3.8 Flash-Next on one RTX PRO 6000 Blackwell (SM120).
Framework-aware code intelligence MCP server for Claude Code and Codex — 90.6% fewer input tokens to review a pull request, median over 60 merged PRs in repos we don't own. 81 languages, 87 frameworks, 100% local.
A browser-based app for designing network architectures visually.
A Frutiger Aero glass dock for Windows 10 and 11. Live window previews, cursor magnification, keyboard search, six themes. Open source, no telemetry.
GPT-6 Astra orchestration with dynamic Sol, Terra, and Luna subagents
A presentation-order collection of 3D prompts for copying and reuse.
Open-source 3D anatomy explorer: 2,234 selectable BodyParts3D meshes, system layers, search, and exploded views.