kaz

Telegram bridges for running Codex CLI and Gemini CLI from a single authorized chat

Pure-Rust streaming speech-to-text + speaker diarization — whisper-rs · sherpa-onnx · tokio async

Shared knowledge base for humans and LLMs on Cloudflare Workers, accessible via MCP, REST API, and Web UI

AI-powered Git commit message generator using Claude Code SDK. Intelligently analyzes code changes and generates contextual commit messages with multi-language support and advanced features.

Project-based Claude CLI session manager with automatic tmux setup and persistence

macOS desktop activity monitor — captures window events & screenshots, reconstructs workflows via LLM (Claude/Gemini/Kimi). Tauri 2 + Rust + React PoC.

kotoba (言葉) - Natural language web testing tool with LLM. Write tests in English/Japanese/Chinese.

AI-powered Git commit message generator with ultra-fast performance (2.5s). One command for add→commit→push workflow.

Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 60-95% fewer tokens, same answers. Library, proxy, MCP server.

Self-hostable dashboard to govern and track a team's AI coding-agent spend: a LiteLLM proxy in front of every provider, request logs in BigQuery, and cost + budgets + adoption metrics via FastAPI. Local-first, MIT.

Workshop and guide for running multiple Claude Code instances in parallel using Git worktrees, with helper scripts.

UAE / Saudi LLM llamafile — Single-file, no-install executables for Falcon, Jais, ALLaM, and more.

Rust FFI / bindgen patterns — from manual bindings to production-grade async wrappers

No description.

Run Claude Code offline on Mac mini M4 — local Qwen3 via llamafile, OS-level sandbox via Safehouse, zero API cost.

High-speed keyword search with shared memory — Go/PHP/Python/Node.js bindings

Local speaker diarization and transcription pipeline comparison for Japanese multi-speaker audio. WhisperX vs pyannote + faster-whisper on Apple Silicon, no cloud APIs.

Nano-fast LLM guardrails proxy in Rust. CPU-only, offline-first, OpenAI-compatible. Filters prompt injection and PII in microseconds — no GPU, no Python.

HIP-1 token deployment validator and budget calculator for Hyperliquid L1. Real-time parameter validation, budget estimation, MCP integration, and E2E tested.

Operate Hyperliquid HyperCore from the HyperEVM — read via L1Read precompiles, write via CoreWriter, bridge HYPE, all wallet-signed. Live demo + runnable code.

Read-only explorer for Hyperliquid HIP-4 prediction (outcome) markets — built on the Info + Exchange APIs alone, no smart contract, no precompile. Live demo + testnet feasibility probes.

Realtime voice AI assistant — streaming STT → LLM → TTS over WebSocket, with barge-in and live translation. Runs fully local (Whisper + Ollama + Piper) or with cloud providers.

A lightweight, lightning-fast, in-process vector database

Configuration demo for observing a LiteLLM gateway with Grafana, Prometheus, Tempo, and OpenTelemetry.

DevSec Scanner — local-first security scanner: SAST, SCA, IaC, DAST, secret, container-image & supply-chain scans from one CLI or desktop app, unified into one report.

Workspace-based RAG knowledge search built entirely on Cloudflare Workers — hybrid (vector + BM25) retrieval, cross-encoder re-ranking, multilingual embeddings, and cited streaming chat.

Know exactly what your Cloudflare usage costs: free-tier vs paid-plan allowances, per-instance cost drivers, month-end forecasts, budgets and email alerts. Sign in with Cloudflare. Built on the Cloudflare stack.

Experimental macOS app exploring whether local LLMs can triage suspicious processes — heuristic pre-filtering + structured risk reports. Built with Tauri 2, Rust & React. POC, not an EDR/antivirus replacement.

Run a local LLM on iOS & Android from React Native — privacy-first, fully offline on-device AI notes (summarize / translate / proofread / chat) via llama.rn (llama.cpp).

Local audio/video transcription CLI with WhisperX, word-level timestamps, subtitles, and optional diarization.

Expo + React Native (iOS/Android) client for AgentClip: capture text & links, search by keyword or meaning, recall with your MCP-enabled AI agent.

Clip browser text, recall it from any AI — Chrome extension + Cloudflare Workers with MCP, semantic search & per-clip encryption

Kanban board where humans and AI agents collaborate on tasks via MCP and REST, running on Cloudflare Workers.

Hyperliquid address profiler that turns public trading data into a Trader DNA card

No description.

Free, open source crypto trading bot

Unofficial Hyperliquid API SDK written in TypeScript

Opiniated RAG for integrating GenAI in your apps 🧠 Focus on your product rather than the RAG. Easy integration in existing products with customisation! Any LLM: GPT4, Groq, Llama. Any Vectorstore: PGVector, Faiss. Any Files. Anyway you want.

MagicTryOn is a video virtual try-on framework based on a large-scale video diffusion Transformer.

A theoretical reconstruction of the Claude Mythos architecture, built from first principles using the available research literature.

The open-source managed agents platform. Turn coding agents into real teammates — assign tasks, track progress, compound skills.

Skills for Real Engineers. Straight from my .claude directory.

🧠 Tauri 2 experimental flashcard learning app with FSRS spaced repetition & LLM-powered card generation

🙌 OpenHands: Code Less, Make More

No description.

Get cooking with ao - the decentralized computer with infinite threads 🐰 🕳️ 👈

Experimental browser-based CAD/BIM viewer on Cloudflare — view IFC & DXF (Japanese/Shift-JIS support), share and manage drawings. React + Hono + Workers + D1 + R2.

Experimental browser IFC viewer for validating IFC4 files.

Experimental browser-only IFC LOI checker/editor demo

Browser-based BIM-to-CAM experiment for curved concrete wall toolpaths.

Create and share 3D architectural projects.

基于Vue3 + THREE.JS 免费开源的三维引擎及配套编辑器，包含BIM轻量化、CAD解析预览、粒子系统、插件系统等功能。 A free and open-source 3D engine based on Vue3 + THREE.JS and its accompanying editor, featuring functions such as BIM lightweighting, CAD parsing preview, particle system, and plugin system.

Minimal setup guide for running a local AI coding agent with OpenCode + Ollama (Qwen3-Coder) — for lightweight tasks, privacy-sensitive code, and as a fallback when Claude Code hits rate limits.

LEAKED SYSTEM PROMPTS FOR CHATGPT, CLAUDE, GEMINI, GROK, PERPLEXITY, CURSOR, LOVABLE, REPLIT, AND MORE! - AI SYSTEMS TRANSPARENCY FOR ALL! 👐

Local-first Markdown/text to narrated video pipeline with Kokoro TTS, ffmpeg, optional cloud voices, and YouTube publishing.

File Parser optimised for LLM Ingestion with no loss 🧠 Parse PDFs, Docx, PPTx in a format that is ideal for LLMs.

MCP Server for IBKR Client

Open Source Alternative to NotebookLM / Perplexity / Glean, connected to external sources such as search engines (Tavily, Linkup), Slack, Linear, Notion, YouTube, GitHub and more.

No description.

JFK file download for LLM analysis

No description.

No description.

My own open source implementation of OpenAI's new Deep Research agent. Get the same capability without paying $200. You can even tweak the behavior of the agent with adjustable breadth and depth. Run it for 5 min or 5 hours, it'll auto adjust.

No description.

Make websites accessible for AI agents

The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery 🧑‍🔬

The ao component and tools Monorepo - 🐰 🕳️ 👈

No description.

ZerePy an open-source launch-pad for AI agents

This repository contains the Virtual governance ecosystem for Virtual DAO and Virtual-specific DAO Governance part. It also covers how Contribution is being recorded and rewarded.

Virtual Protocol frontend SDK for ReactJS. Contain React components and hook necessary for integrating with Virtual Protocol server. Frameworks other than React can also use the services and helper functions to speed up integration.

No description.

Open source software that helps you create and deploy high-frequency crypto trading bots

No description.

Middleware that standardizes DEX API endpoints on different blockchain networks

An experimental open-source attempt to make GPT-4 fully autonomous.

No description.

operation toolkit for Avalanche nodes

No description.

Launch your own EVM as an Avalanche Subnet

No description.

No description.

WeaveDB - Decentralized NoSQL Database powered by Arweave

No description.

No description.

:chains: A Framework for Building High Value Public Blockchains :sparkles:

Opinionated Framework for Building Hyper-Scalable Blockchains on Avalanche

The Context Layer of the Decentralized Web

React Hooks for Ethereum

🏗 forkable Ethereum dev stack focused on fast product iterations

Hardhat is a development environment to compile, deploy, test, and debug your Ethereum software. Get Solidity stack traces & console.log.