Havuz

面向 AI Agent 的高并发、低延迟上下文治理与 Prompt Cache 守护网关 (Tree-sitter AST / Tool Delta / 50%~80% Token 削减)

Loopcut: an AI agent that works inside Blender. A fork of Blender (GPL).

Bilingual (中文+EN) ML / LLM / diffusion / agent interview cheat sheets for AI 秋招 — generated by ARIS /interview-cheatsheet, rendered by /rend

LLM agent 的有界自纠错长期记忆层：把连续、含矛盾的项目交互日志蒸馏成有界、可自更新的记忆库。Bounded, self-correcting long-term memory for LLM agents — candidate pool + value dynamic

llama.cpp fork for significantly improved performance on Ampere generation Nvidia cards (especially RTX 3090 / 3090 Ti): example: 110+ tok/s

nopus, coding-agent yanıtlarındaki okunabilirliği deterministik metin ölçümleriyle değerlendirip gerekirse bir kez daha açık yazılmasını isteyen bir eklentidir.

Neden yükseliyor · 43 günde 296 yıldıza ulaşması, coding-agent araçlarına dönük ilginin ve yanıt kalitesi sorununa yönelik merakın göstergesi olabilir; ancak yükselişin nedenini doğrulayacak veri yok. Son 28 günde commit ve farklı yazar sayısının sıfır olması güncel geliştirme etkinliği konusunda kanıt sunmuyor.

Claw-Eval is an evaluation harness for evaluating LLM as agents. All tasks verified by humans.

安卓AI聊天端:开箱即聊，无损导入酒馆卡，缓存强省用量，Jev 决策，专属RP模式，防空回，主动信息，前端卡支持，高定制主题，QQbot，AI群聊，插件系统，强兼容中转站。手机移动端原生支持，求 Star 收藏⭐Android AI Chatbox: lossless Silly

Local decision engine with typed, calibrated outputs (Noul/Choice/Score): cheap→strong cascade, ECE/Brier calibration, SystemOne protocol in

21 writing rules for AI coding and writing agents. Drop-in for Claude Code, Codex, Copilot, Cursor, and Aider, so their output reads like a

Open-source directory of agent-bot prompts for Grok Bot, Rakazo, and any agent — botdirectory.ai

dsh-sieve: context engineering & token optimization plugin for DeepSeek Harness (DSH) — tool output filtering, context pruning, progressive

The agent flow system for token maxxing.

From teacher to tiles — a from-scratch LLM distillation & serving engine: custom Triton/CUDA kernels, FSDP distillation, paged-KV continuous

An open-source Cotypist with macOS system wide AI autocomplete

A 14M-parameter LLM keeps a virtual fish tank alive on an $8 ESP32-S3 — distilled from a 26B teacher into 7.56 MB, no network.

From ticket to reviewed pull request. Free and open-source, on your machine.

Transformers for decision models: fine-tune, load and serve calibrated decsion-model on any Hugging Face model.

Uncensored & abliterated AI models (GLM-5.3, DeepSeek V4, Qwen3) + Claude, GPT, Kimi, Gemini — one OpenAI/Anthropic-compatible API key. Docs

A fast browser agent: Jev picks each action from what is on the page, an LLM reads and plans, and every claim in an answer cites a quote fro

Find zero-days while you sleep. DeepZero is an automated vulnerability research framework that parses, decompiles, and analyzes thousands of

A minimal modern terminal for AI coding experience — Sidebar workspaces; horizontal / vertical split panes; one-click agent launch; per-agen

LLM-maintained personal knowledge base for Obsidian. Based on Andrej Karpathy's LLM Wiki pattern.

Autonomous self-evolving agents. Vision-grounded layered memory and self-written skills for LLM agents that operate your computer.

Quicksilver, toplu sınıflandırma ve arama işlerini Jev modeline devrederek Claude Code'un daha az token kullanmasını amaçlayan bir skill ve plugin'dir.

Neden yükseliyor · README, 12 görevlik benchmark'ta token kullanımında medyan %82 azalma ve bazı görevlerde hız artışı iddia ediyor; ölçülebilir tasarruf vaadi dikkat çekici. Ancak bu sonuçlar repo'nun kendi benchmark'ına dayanıyor ve bağımsız olarak doğrulanmış değil. Proje yalnızca 2 günlük olduğundan 66 yıldız erken ilgi göstergesi; 2 commit ve 1 farklı yazar ise henüz sınırlı geliştirme kanıtı sunuyor.

Open-source local AI workspace — advancing on-device inference.

The ultimate collection of APIs for building autonomous AI agents — 2,036 production-ready APIs across Agents, AI Models, and MCP Servers. S

30KB Markdown-native WYSIWYG editor for React, built for AI streaming, human editing, huge files, macOS, Web, and agent workflows.

A multi-agent research system for the humanities and social sciences.

Aiden — an autonomous AI agent and work engine built solo. It can operate your browser, terminal, files, apps, APIs, skills and tools, remem

MCP server and CLI that crawls websites for AI agents: finds the answer page 2x more often than Crawl4AI, 10x cheaper per answer than lettin

Harness-oriented agent system framework for production-grade LLM agent applications

A local-first desktop app for running a team of AI agents on Claude Code, Codex, Cursor, Cline and more, on your own computer.

Ryza Chat, kullanıcının seçtiği LLM ve TTS servislerine bağlanan, gerçek zamanlı 2D avatarlı bir AI companion istemcisidir.

Neden yükseliyor · 26 günlük repo için 55 commit ve 4 farklı yazar aktif geliştirmeye işaret ediyor; 193 yıldız da ilk dönemde dikkat çektiğini gösteriyor. Bununla birlikte yıldız ve README, kullanıcı benimsemesini veya ürün kalitesini tek başına kanıtlamaz.

Estrella Bot · QQ 群 AI 群友机器人。把 QQ 小号接给任意 OpenAI 兼容模型（本机 / DeepSeek / 智谱 / 通义），让 AI 像群友一样插话、接梗、潜水。OneBot v11 + NapCat，零构建网页控制台。Apache-2.0

Grasp helps you manage multiple projects and parallel coding agents in one visual workflow, so humans can understand faster and ship more.

wanman is an open-source agent matrix runtime inspired by Japanese one-man trains. It lets human users step back into an observer role while

A prior-art search for your code ideas — has this dev tool already been shipped?

The open-source operating system for physical AI.

AI 应用开发、AI 编程实战与面试指南，涵盖 LLM、Agent、RAG、MCP、Claude Code、Codex 等核心技术与工程实践。

gap-trap, coding agent kullanan repolara projeye özel kalite kuralları ve bu kuralları CI veya test sürecinde denetleyen gate'ler ekleyen bir agent skill'dir.

Neden yükseliyor · 14 günde 180 yıldız ve son 28 günde 34 commit, yeni projeye yönelik belirgin ilgiyi ve geliştirme hareketliliğini gösteriyor. Ancak proje çok genç ve farklı yazar sayısı 2; bu veriler yaygın benimsemeyi veya uzun vadeli sürdürülebilirliği kanıtlamıyor.

🐚 The AI assistant that just works. One self-hosted app for every model you already pay for: Claude Code, Codex, Copilot, Gemini, Ollama and

DeepSeek Harness gate - per-turn persistent-operation counter: remind, instruct the model to stop, then ask for confirmation. Steps aside un

基于 DeepSeek Harness（DSH）的稳定桌面端，集成Git、内置浏览器与记忆功能 | DeepSeek Harness desktop GUI with local workspaces, Git, browser and memory.

Your car as a chat-room agent: Raspberry Pi 5 + dashcam + local AI. CodeWatch's sibling for the garage.

⚡ Native MLX Swift LLM inference server for Apple Silicon. OpenAI-compatible API, SSD streaming for 100B+ MoE models, TurboQuant KV cache co

MathCode: A Frontier Mathematical Coding Agent

让 AI 住在你 MacBook 的刘海里 · 零依赖开箱即用 · 多引擎并行的桌面 AI 伴侣（Swift 6 / SwiftUI / macOS 14+）

The local-first LLM Wiki: open-source knowledge graph builder, RAG knowledge base, and agent memory store. Built on Andrej Karpathy's patter

A free, open-source desktop app for getting work done with AI. It shows you what it would remember before it saves anything.