Universal LLM router for AI coding tools. Works with Claude Code, Cursor, Codex, Gemini CLI, Copilot and more. Free-first fallback chain cuts costs 35-80%.
- ai-routing
- anthropic
- claude
- claude-code
132,241 visible repositories in the index
9,361–9,384 of 10,059 · Most stars
Universal LLM router for AI coding tools. Works with Claude Code, Cursor, Codex, Gemini CLI, Copilot and more. Free-first fallback chain cuts costs 35-80%.
Face Recognition on NIST FRVT Top Ranked, Face Detection, Face Matching, Face Analysis, Face Sentiment, Face Alignment, Face Identification && Face Verification && Face Representation; Face Reconstruction; Face Tracking; Face Super-Resolution on Windows
Axon/Myelin segmentation using Deep Learning
Rule-based personalisation for Wagtail CMS
TranslatedMcDonald's points asset steward: filter out fake listings · rank by value for money · lottery advisor
A sklearn-compatible Python implementation of Multifactor Dimensionality Reduction (MDR) for feature construction.
Turn intent into work. Open-source Desktop/Web agent harness with long-horizon memory, live evaluation and recoverable batches.
TranslatedMultimodal retrieval in practice: intent recognition, prompt engineering, Qwen LoRA, EmbeddingGemma 2, and InfoNCE contrastive learning. Includes real demos, integration of your own data, training and evaluation, and a learning guide to RAG retrieval components.
TranslatedWhat to eat when working overtime: a WorkBuddy Skill for McDonald's MCP group meal planning, combo allocation, order-change repricing, and exact bill splitting
Writeups of Capture The Flag Competitions
ReAlign models instantly, on demand. Per-request alignment for verified users and channels, or Always On. Use directly or with Hermes, OpenClaw and custom harnesses.
Command line utility and Python package to ease the (un)mounting of forensic disk images
Generate the smallest test dataset that provably catches broken data-mapping rules in ETL pipelines.
RouteAlign
Produce watertight polygonal vector maps by tracing raster images
TranslatedMcDonald's meal pairing decision Skill | Computes the top three optimal choices out of 1881 combinations, and tells you why they are not the others. Calorie control, dietary restrictions, and max-min fair allocation for group orders; the LLM does not take part in calculating money, and every number is traceable.
TranslatedA McDonald's ordering solver under the dual constraints of budget x calories | McDonald's MCP Skill
Terminal CLI to search and download ROMs, ISOs and Switch games. Interactive Textual UI, cross-platform catalog, resumable downloads.
MiniMax H3 pictures, talking video, extensions and TTS on your own Google Colab GPU — a skill for Claude / Codex / Hermes
DeepSeek Harness Desktop 插件,让 AI Agent 用自然语言构建、检查和维护可编辑的 SideFX Houdini 工程,支持程序化建模、HDA、场景诊断与渲染。 | A DeepSeek Harness Desktop plugin that lets AI agents build, inspect, and maintain editable SideFX Houdini projects with natural language, including procedural modeling, HDAs, scene diagnostics, and rendering.
Portable AIX Protocol Agent Skill
Your personal AI agent
Translated麦麦健身餐规划师 (Maimai Fitness Meal Planner, "McDonald's fitness meal planner"): Going to the gym is not as good as Go To ᗰ! This is a McDonald's ordering Skill that is perfect for gym-going foodies, boys and girls alike! Whether you are in a muscle-building phase or a fat-loss phase, just tell it the calories, protein and budget you want for this meal, and it will automatically calculate the nutrition of the items and recommend money-saving combinations. Even better, its nutrition calculation is precise enough to account for removing sauces or ingredients, and it gives suggestions for splitting a meal across several sittings, so you control your calorie intake while reducing food waste. It builds training-plan-style ordering plans based on the McDonald's China MCP.
Run Qwen3.8-27B locally on Windows with one NVIDIA RTX GPU: chat and an OpenAI-compatible API. On an RTX 5090 the fastest model averaged 2,298 tok/s on our own JSONL edit task (one request at a time, thinking off, v0.1.4).