Search

Pi extension for Cloudflare Workers AI — chat, image generation, TTS, and speech-to-text over the free tier

Shared interactive tools runtime for Tomako-FE workspace dialogs and Tomako-SEO tool pages

Scan a website with agent-readiness scanners (afdocs, Vercel Is Agentic, Ora), get a unified scorecard, auto-fix supported issues (Astro + Starlight, or plain Astro, + Cloudflare Pages, Netlify, or GitLab Pages; Next.js + Vercel or Netlify; Jekyll + GitHu

Local context routing and MCP setup for OneShot Copilot workflows

插件按钮聚合器（高度可扩展）：把三方插件散乱按钮聚合到统一入口——SSiD 标题栏 / web 悬浮球；载体+驻场 LLM 范式——任何人都能把自己环境里任何插件的按钮动态聚合（LLM 一键注册、右键删除=注销，纯数据零代码，不动插件源码）

OpenCode provider wrapping the Codex app-server agent

Production-grade AI Frontend Orchestrator for React, Next.js, and TypeScript applications. Governs Figma-to-Code mapping, context budgets, security boundaries, visual QA, and autonomous repair.

Extract PyPI metadata — version, author, license, homepage, download counts — for the core AI/ML Python stack: transformers, torch, langchain, openai, anthropic. Official PyPI JSON API pipeline. $1.00/1K, no auth required.

A modern node:stream polyfill

[English](README.md) | 中文

Read a page, a screenshot, or a video as text. Default path never sends pixels off the machine.

Post-processing effects for any SVG. Scanlines, bloom, glitch, halftone and more — in, out, still an SVG.

Automatic multi-client (OpenCode, Claude Code, Hermes), multi-protocol, multi-provider local AI gateway and prompt engineering studio.

Framework-agnostic prompt editor web component and TypeScript client SDK for the UTD Prompt Manager client API

Production-grade software engineering lifecycle skills for coding agents.

Part of PAC: @projectpac/flow-template.

Composer 'Web Search' toggle for the DeepSeek Harness Web UI. When on, the agent is forced (via a mandatory system-prompt directive) to call web_search before answering.

A visual editor for Open Semantic Interchange (OSI) semantic models

Choose the right AI model for the task: ranked, explainable model recommendations with cost, speed, task fit, and provider allowance.

Ainize node — serves a base model whose conditional memory is extended by knowledge patches; teaches, verifies, publishes and trades patches on the AIN ledger