Turn a phone voice note into a studio mic voice. Local and free, by @justbyamine

Redline: code review inside the Herdr app. Review GitLab MRs and GitHub PRs in a git worktree next to an AI agent: inline comments, pending reviews, council AI findings, pipelines, approve and merge.

Open-source desktop AI trading assistant for market research: chat beside live charts, price alerts, local backtests, bring your own model.

One Go binary that lets a client talk to an AI contact the way it would talk to a person: JIT threads, JIT prompt cache, JIT REPL environments over one append-only log.

Analyze jailbroken iOS and macOS apps, then build and verify independent implementations of selected behavior.

Research internship cold-email tool: finds professors matching your resume, writes personalised emails, sends them paced from your inbox, tracks replies. Free, open source. By Avinash Chandra Gupta.

FALSIFY · 证伪 — 先证伪，再决定。A portable reasoning kernel: evidence layering, falsifiers, and a tamper-evident forecast ledger for any AI agent.

AI anticheat for Minecraft servers: neural networks catch KillAura, aim assist, triggerbot and reach. Paper, Spigot, Folia 1.16.5–1.21.

Policy-adherent, auditable reasoning for AI Agents.

网文平台资源配置决策助手：GraphRAG 两阶段检索 + ReAct Agent + 规则引擎兜底，方案合法率 60%→100%

MCP server for Apple Developer Documentation: API reference, Human Interface Guidelines, App Store Connect Help, sample code and 1,700 WWDC sessions with transcripts and code. A fork of Apple Docs MCP.

Claude.ai in your terminal: streaming chat, OpenAI proxy, MCP server. No API key.

Python voice AI assistant: microphone STT, Google Gemini answers, offline TTS with pyttsx3. Clean env-based setup.

Host-side scheduler and transfer-path optimization in SGLang on H100 with C++ descriptors, Triton and pinned memory

AI portrait SaaS (Laravel 11, Livewire, Filament, Stripe, fal.ai) — run in production, released as open source after closing

Tying up what The Sims 4 leaves dangling

AI security training: hands-on labs, tools and courses for LLM, RAG and agent security. Maintained by Pwned Labs.

Claude skill + plugin that picks the right model and effort for every task: Opus plans, Haiku/Sonnet build, every step is verified, and it escalates only on failure. Save tokens without losing quality.

跟 AI 下棋、看客观数据、听本地模型讲棋的围棋工具（Electron + KataGo + LoGos）

⚖️ Kadı — a calibrated judge in front of your frontier LLM. Routine decisions answered locally with a conformal-prediction guarantee (≤ α disagreement with the oracle); everything else escalated and learned from. Zero-dependency Python.

train your own chatbot from scratch on a mac

风插 2.0 — DSH 推理语域锚定预设：两个 agent preset（风插标准/风插 PTC）在 runtime-context 快照头尾各注入一次形式化约束，使推理以第一人称复数展开。定性=硬指令约束（非吸引子）；未证明能力提升，未证明无负面作用。Apache-2.0

AI가 청중에 맞는 HTML 발표자료와 발표 노트를 만들고 Chromium 기반 시각 검증까지 수행하도록 돕는 스킬·플러그인

A domain-specific medical Retrieval-Augmented Generation (RAG) system that answers health-related questions strictly based on retrieved evidence from a local vector database.

Python crypto scanner: finds coins in quiet accumulation, back tested on 961 samples, rebuilt from the evidence, automated with Telegram alerts.