Browse agent skills

VinaX: your AI coding agent for the terminal, powered by Groq, OpenRouter and NVIDIA models.

On-device text-to-speech in the browser: Phonon compiled to WebAssembly, streaming 24 kHz speech from a worker.

YAML parsing and document loading with local and remote references, no dependencies.

Pi Agent extension: synchronize Ollama models into pi's models.json so they appear in the /model picker

Self-contained System 1 decision engine: multilingual INT8 transformer served by a zero-dependency native binary (or the bundled pure-Rust WASM engine). TypeSafe Jev wire-compatible (/v1/systemone). Runs offline in Node.js, Bun and browsers.

Local-first semantic slicing for long ecommerce images, with previewable cut plans for Codex and SealSeek.

AI-first rich-text editor SDK on Tiptap: streaming AI writing that lands as real document nodes, AI-generated formulas and diagrams, plus toolbar, bubble menus, tables, code, images, 16 languages and session replay. React UI today, framework-agnostic core

Stealth mode: Applies various techniques to make detection of headless puppeteer harder.

Qriton CLI is a terminal coding agent that runs on infrastructure you control. Local models or any major model API, with a memory record, HLM, and Energy Language.

A curated, auto-synced collection of agent skills (design, animation, dev workflow) with a CLI to install them into any harness: Claude Code, OpenCode, Cursor, Codex, Gemini.

Drop-in AI chatbot for Next.js. Install, add your Groq API key, done.

Run LLMs in the browser — GPU or CPU, same API. Tool calling, embeddings, RAG, offline knowledge bundles, and metrics. Hooks-style TypeScript over Transformers.js (injectable).

On-device Quran recitation recognition for React Native / Expo: onnxruntime runner, recognition pipeline and a lifecycle-safe microphone hook.

Minimal Expo module that streams mono float32 microphone PCM (e.g. 16 kHz) to JS in fixed-size chunks, for on-device speech recognition.

On-device Quran recitation recognition core: CTC decoding, verse identification and a conservative trust gate. Runtime-agnostic (inject your ONNX runner).

TypeScript SDK translated from ccwork-chat-sdk-go

ZeroSub: use several Claude and ChatGPT (Codex) subscriptions side by side in Paseo. Add accounts in-app, pick one per agent, and switch automatically when an account hits its limit.

Local bridge server for MarkdownPad — enables Claude AI assistant from the web app

Resolve and cache Hugging Face Hub files: ~/.cache/huggingface layout on Node/Bun, Cache API/OPFS in browsers

Orchvia engine: one local scheduler with durable SQLite state for Claude Code and Codex agents.

A small, type-first CLI framework for Node.js — commander.js-like ergonomics, fully typed.

Every frame drawn in JavaScript — the Opus 5.5-style animation technique as agent skills: hand-drawn canvas animation, zero image assets, code-synthesized soundtrack, A/V sync check.

Typed HTTP client with CRUD services, plugins and injectable transports.

Local speaker embedding, enrollment, and identification with Sherpa-ONNX WASM

Dibbla Agents SDK for TypeScript: workers that serve functions, jobs and capability providers to the Dibbla workflow engine

n8n nodes for the ContactWise messaging API: send DLT-compliant SMS to Indian mobile numbers

Framework-free bracket data linking and prediction resolution for the Ethlete SDK.

Browser HTTP client for map-store /map-store/v1 (Memorix estate via map-store only)

Local speech-to-text for Pi

输入关键词、快手账号主页链接，自动拉取垂类作品、账号作品并解析无水印下载链接，支持一键批量下载到本地。支持三大能力：(1) 关键词搜索视频，可按点赞数、发布时间、视频时长筛选排序；(2) 达人作品抓取，按主页链接获取公开作品列表；(3) 视频评论分析，按视频链接获取评论内容与互动数据。