Code agents

OpenAI-compatible API gateway for your Cursor subscription. Point any OpenAI client at it — streaming, tool calling, sessions, 80+ models. Self-hosted, with web dashboard, CLI, autostart & Hermes Agent addon.

Clean My Codex is an open-source disk space analyzer and safe cleaner for OpenAI Codex Desktop on macOS and Windows. Clean old sessions, generated assets, plugins, and workspace output.

AI-powered codebase Q&A to explore repositories, docs, and project insights

自動化 OpenAI 帳號註冊與 AI API 代理管理工具 | Automated OpenAI account registration and AI API proxy management tool

🧠 Mnema — Long-term memory for AI via MCP × Vector DB. Pluggable backends, hybrid search, Auto Dream consolidation, CLI + REST API + SDK.

Smart git diff context for LLMs and AI agents: budgeted fragment selection, deterministic, MCP server included

Keep Codex Desktop conversation history visible when switching subscription and OpenAI-compatible API profiles. Local metadata repair, backups, and private credentials.

AI agents that work, learn and are governed like an organisation. One installation, one database, one interface: org chart, budgets reserved before every call, approvals, encrypted secrets, audit, memories and skills, routines, MCP and Telegram.

AI Governance Platform - LLM Gateway, MCP Gateway, A2A Gateway.

Self-hosted — a unified workspace where humans, AI personas, and autonomous agents collaborate on complex problems through a single, shared reality

Lightweight Windows tray monitor for Codex 5-hour and 7-day usage, remaining quota, and reset times. 本地、轻量、隐私友好的 Codex 用量监控工具。

Background dictation for Windows that survives switching language mid-sentence. Hold Ctrl+Alt, talk, and the text lands in whatever you were typing in.

DeepSeek for free — no subscriptions, no API bills, no middlemen. Private, powerful and ready for your favorite apps and coding agents.

Note-grounded AI chat, reviewable edits and image workflows in Obsidian. Co-developed with Prof. Changhyun Ahn. By CMDSPACE.

Enterprise-grade, high-performance LLM serving on Kubernetes. OpenAI-compatible inference powered by vLLM and OpenVINO™ Model Server behind a unified AI gateway, with NUMA-aware CPU optimization for Intel® Xeon® - production model serving, routing, and security built in.

Demo application to chat with a book (using LangChain and Open AI GPT LLM)

AI 辅助接口自动化测试平台，支持多环境、链式变量、智能用例生成与 HTML 测试报告。

A tiny macOS menu bar app for OpenAI Codex usage limits

Awesome list for Claude API usage: Console, docs, cost and usage reporting, pricing, cloud partners and alternatives.

GLAW — open-source virtual law firm AI-agent system: 183 skills, 181 commands, hard-gated matters, MCP/API intake, and X402 paid-agent charge matrix with Conway-style agent survival. Attorney work-product for licensed review, not legal advice.

MCP server for Zotero 10+. Read and write your library entirely offline through Zotero's native local API. No plugin, no API key, no cloud.

Turn a Claude Code subscription into an OpenAI-compatible API: Self-hosted LiteLLM gateway for chat completions and agentic coding sessions.

Claude Code skills for content creators. Cut silence from video, normalize audio, remove filler words.

Curated analytical benchmark directory, latency scanner, and client config generator for free and trial AI API relays.

The open-source Claude skill for UI/UX design. 84 UI styles, 192 color palettes, 74 font pairings, 99 UX guidelines & 161 industry rules across 22 stacks. Works with Claude Code, Cursor & Windsurf.

Talk to Claude Code on your own machine, from your phone, over Telegram. No tunnel, no webhook, no cloud relay.

Run Claude Code natively inside Oh My Pi — AskClaude tool + switchable 1M/200K context window

Offline HTML viewer for Claude Code JSONL session transcripts. Six UI languages, light/dark theme, no backend, no CDN.

A Claude Code plugin that turns the words you reach for while working into your own study material. Instead of someone else's word list, you learn the language your real projects speak, and you review it right where you work. Everything stays on your computer. Build your own Quizlet or Duolingo out of your own data.

Copy-paste Claude Code hooks that actually do something — auto-format, block dangerous commands, test gates, notifications. Dependency-free, current schema.

Lightweight, fast OpenAI-compatible AI router in Rust. Prefix routing, fallback combos, token saver, usage metering, embedded dashboard. One binary, no runtime, no external DB.

Multi-agent AI assistant for ICU clinical teams, retrieves evidence from medical literature, analyses chest X-rays, brain MRIs & skin lesions, and answers clinical questions using RAG, LangGraph, and GPT-4o.

Local code retrieval for codebases too big to grep

制造业 RAG 与设备运维智能体：支持企业知识问答、故障诊断、工单生成及 FastAPI 系统集成。

Local-first, battle-tested, memory that never disappears. Hybrid vector search + knowledge graph + full-text retrieval for AI agents.

Secure, local-first document-to-Markdown conversion. Zero-trust input sanitization, multi-column PDF layout, LaTeX math, PII masking. No cloud dependencies.

Fully local MCP retrieval tool for coding agents. 161M encoder, ~1s per tool call, no API, no LLM in the query path. No file reads: line numbers, callers, callees and exact call sites, packed to a token budget. 2-4x the file coverage of BM25 and TF-IDF on SWE-Explore, and tops the paper's own sub-1B code model on CORE-Bench.

Local-first persistent memory plugin for opencode: three-layer memory (long-term knowledge / episodic summary / working memory) with SQLite + Chinese-aware retrieval (RAG), fully local.

面向计算机考研 408 的本地优先智能学习助手：以 RAG 全链路可溯源问答，内置四科知识点与 1175 道选择题，AI 出题即时判题；自带 200 题量化评测体系，Docker 一键部署，学习数据全程本地运行。

Local memory + semantic search for AI agents. Indexes code and documents on your machine, queryable via MCP. Private, offline, fast. macOS arm64 + Linux x86_64.

Local-first PDF reader for researchers: notes anchored to the page, AI answers that cite it. Runs offline with Ollama.

Self-hosted AI audio analysis & knowledge workspace for transcription, speaker/emotion review, business insights, and RAG.

Local-first personal AGI GeoSentinel node — any OpenAI-compatible LLM plus dual-lane FAISS memory, offline voice, live global telemetry, sandboxed tools, a Discord bridge and an MCP server. No cloud required.

MCP context engine for AI coding agents: a local code graph answers "which files does this task need?" — folded evidence packets instead of whole files. Recall 1.000 on repos it was never tuned on, 45% fewer context tokens in a real Claude Code session. Rust, local-first, MIT.

Desktop GUI Headless Local-first AI coding agent in Rust (axum, tokio, rmcp, portable-pty) behind HTTP + SSE, with an Angular 20 zoneless client, Tauri 2 desktop shell(and mobile in near future). OpenAI/Anthropic/Z.ai/Ollama + more, MCP bridge, permission engine, PTY terminal.

ZYRAXON AI — The World's Most Advanced Desktop AI Agent.

在飞书妙搭（apaas）上用 AI Agent 开发的避坑手册 + harness。83 条实测踩坑，一条命令接入。

⚡ Governed economic agency for autonomous agents on Tether's WDK: an authorization plane, an isolated signer that re-derives every call, one agent object behind them. Testnets only, by decision — no flag widens it. Apache-2.0 — npm install && npm test. The mirror is public; the plane behind it is not.

End-to-end Claude Code skill for job applications. 8 phases, 2 sub-skills, originally built for UK AI startup hunt, generalised. Bilingual EN/中文.

FTShare-powered terminal financial visualization canvases for Codex and OpenCode｜面向 AI Agent 的交互式 K 线、行情、新闻与数据画布

Free, open-source cross-platform desktop app for the Pi coding agent on Windows, macOS, and Linux. Projects, conversations, streamed tool activity, diffs, model controls, and graphical extensions in one native workspace.

A Cockpit plugin that turns the web console into a full file manager — browse, edit, preview, search, terminals, and Git — with integrated Claude & Codex AI CLI tabs showing a live git diff, plus fstab/mounts, SMB & NFS shares, and a GRUB editor.

Pi harness as a modern app, with UI, web search, subagents, /goal, and more, all installed in one command

Agent skill that distills AI study chats into a causal-chain learning wiki — mechanical checks, a regression eval gate, evidence-driven self-iteration (RSI). DSH / Codex / Claude Code. ｜ 把 AI 学习问答沉淀为因果链学习 Wiki：机械校验 + 回归门禁 + 证据驱动的自迭代。

🌐 Drop-in skills that teach Claude Code, Cursor & Codex CLI how to automate browsers. Playwright, Selenium, Puppeteer, Web Scraping, AdsPower & more.

Product Design Agent Skills for Claude Code & Antigravity IDE

Safe, Speedy, Self-evolving coding agent.

Local-first AI-to-3D pipeline: LLM planning, validated Blender generation behind an MCP tool boundary, read-only vision review, semantic motion, and GLB/MP4 export with a FastAPI + Three.js studio.

万物皆是插件，万物皆可自定义——基于DeepSeek Harness的安全测试工作台，提示词、工具、MCP、技能、知识库、WebShell均可模块化编排，覆盖渗透测试与代码审计。

Windows computer-use MCP server (26 tools) with a human-in-the-loop approval gate: physical-input-only dialogs, signed guard switches, deny/launch lists, hash-chained audit log. 70+ unit tests, green CI.

An AI agent that answers your phone calls, resolves the routine ones, and pulls you into the same live call when it needs you.

Open-source AI SRE agent that powers the native OpenObserve AI Assistant chat box — read-first grounding, SQL/VRL/PromQL generation, and confirmation-gated actions. Zero front-end changes.

用自然语言查询、分析项目数据的 AI Agent

Local-first AI browser automation with Playwright. Use OpenAI, Grok, DeepSeek, Anthropic and compatible APIs—or ChatGPT/Claude Code subscriptions. Approve site access, watch each action, and replay proven workflows with zero model calls.

AI-directed job search system: scans target companies for matching roles and generates tailored, no-fabrication resumes/cover letters. Fork it for your own search.

A visual AI Agent workflow platform with LLM, RAG, MCP, tool calling and streaming execution.

Open-source, self-hosted alternative to Lovable, Replit and Base44. Create full stack web apps on your self hosted instances for free.

AI-Powered Code Security Analyzer. Pluto is a powerful CLI tool that uses AI to detect security vulnerabilities in your code.

Model-agnostic DevOps AI agent harness: policy, approvals, audit and rollback enforced outside the LLM. Works with Claude Code, OpenCode, Cursor, Copilot and any MCP client.

Provider-agnostic coding-agent runtime and CLI for Elixir and the BEAM

Experimental open-source AI coding agent in Rust

Evidence-grounded, auditable Codex skill for SCI review research and writing.

Moved to https://github.com/nano-muse/nanoMuse — archive of OpenMuse 0.1.0–0.6.0

BubbleBuddy is a friendly Discord AI companion powered by Pi, Effect, and discord.js

Codespaces-ready starter for embedding a live three.ws 3D AI agent in any site

AI_CONTEXT: a way to share durable project context between any AI agents and humans using the repository itself as the continuity layer.

通用电商店铺首页设计 agent skill：先问出图尺寸 → 概念驱动差异化 → 生成可视化编辑器，导出任意平台尺寸图片。适配各类 coding agent。Platform-agnostic e-commerce store-homepage design skill for AI coding agents.

Local-first autonomous AI agent terminal with a native Rust runtime, SQLite persistence, SAFE execution, and a React/Ink TUI.

AI-native terminal & SSH client — Harness launches DeepSeek / CodEx / Claude agents on PowerShell 7 in TUI or embedded Web UI, plus SFTP, serial/Telnet and a built-in MCP server.

P2P multi-agent scenario system for the Pi coding agent: comms-net hub (HTTP/SSE, cross-device) + one-command scenario launcher. Agents that build agents.🌱

Research-backed Fiverr gig optimizer — a Claude Code skill with deterministic competition + pricing scoring (no ML, no guessed market numbers).

Tools by AIs for AIs

IMA (ima.qq.com) knowledge base MCP server - search, browse, read notes, and add content to your IMA knowledge bases from any MCP-compatible agent (Claude Desktop, Cursor, Cline, VS Code, OpenClaw, WorkBuddy). MIT licensed.

Drop a card on the board, pick Claude Code, Codex CLI, OpenCode or the Hermes agent, and watch it work in its own git branch and worktree — then review, test and merge from one screen.

Local CLI/MCP bridge that lets Codex delegate compact review, research, and planning tasks to ChatGPT Web.

The last mile of the academic figure toolchain — a Claude Code skill that draws editable, paper-exact vector figures directly in Figma via MCP. 学术绘图工具链的最后一公里

A control plane for trustworthy AI agents: write product judgment once, then enforce it through typed MCP tools, git gates, demolition-first stories, evidence-judged checkboxes, and zero-trust monitoring. Ships with Claude Code and Cursor plugins, Agent Skills, and a FastMCP server.

Claude Code mode switcher for GSD, gstack, Superpowers, Oh My IPD, OMC, and OpenSpec workflows.

Terminal multiplexer and agentic control plane with tailscale enabled. optimized for mobile development

Share sandboxed HTML previews with public-unlisted or access-key private links. Includes a remote MCP server for AI agents.

Universal cross-platform repo-local workflow engine for AI agents: context-as-code, reusable skills, and one-command install for Cursor, Claude Code, OpenAI Codex CLI, and OpenClaw.

A governized pipeline fit for Claude Code to deliver quality output comparable to an established team of software and web engineers within modern standards

AI-Powered Inclusive Education Referral Continuity & Statutory Assessment Compliance Platform (India)

Self-hosted durable execution engine with built-in observability for AI agent runs

iOS application development skills including develop, ASO etc

Agent skills for design taste and visual judgment. Hierarchy, typography, color, spacing, interaction quality, and the invisible decisions that separate good design from great design

20 AI agents for video creation, copywriting, landing pages, and ads with Remotion Studio and Agentation visual feedback

Deterministic agent loops for AI coding CLIs (Claude Code, Codex, Cursor, Kimi): worker + adversarial verifier agents, hard quality gates, automatic retries, human escalation. File-based, zero dependencies, one-click on/off.

FEMO插件版 — 可接入dsh的多智能体引擎，差不多能做所有事：我已经用femo和我的主agent玩了好几局狼人杀了，他抽到狼他还刀我！ 我在b站的宣传视频也是用femo剪的。 coding的时候，特别难找的bug、代码重构，我都会用femo来做。 我甚至用femo模式给我的主agent养了只宠物，她很开心。