Open Design: Open-Source Alternative to Claude Design Runs on Your Local CLI Agents

✍️ OpenClawRadar📅 Published: May 2, 2026🔗 Source
Open Design: Open-Source Alternative to Claude Design Runs on Your Local CLI Agents
Ad

Open Design (OD) is an open-source alternative to Anthropic's Claude Design, released 2026-04-17 on Opus 4.7. Unlike the closed-source, paid-only, cloud-only original, OD is local-first, web-deployable, and BYOK at every layer. It auto-detects 11 coding-agent CLIs on your PATH—Claude Code, Codex, Cursor Agent, Gemini CLI, OpenCode, Qwen, GitHub Copilot CLI, Hermes, Kimi, Pi, Kiro—and wires them into a skill-driven design workflow. No CLI? An OpenAI-compatible BYOK proxy works the same loop without the spawn.

Key Features

  • 72 brand-grade Design Systems and 31 composable Skills (including bundled skills/guizang-ppt-skill from op7418/guizang-ppt-skill for magazine-style decks with WebGL hero and P0/P1/P2 checklists)
  • 5-step brand-asset protocol from alchaincyf/huashu-design, with anti-AI-slop checklist and 5-dimensional self-critique; direction picker offers 5 schools × 20 design philosophies
  • Live TodoWrite plan streams into the UI; agent builds a real on-disk project folder with seed template, layout library, and self-check checklist
  • Sandboxed iframe preview (vendored React 18 + Babel) with live agent panel showing todos, tool calls, and interruptible generation
  • Export to HTML, PDF, PPTX, ZIP, Markdown
  • Runs locally via pnpm tools-dev; web layer deployable to Vercel
Ad

Architecture

OD stands on four open-source projects: huashu-design (design philosophy), guizang-ppt-skill (deck mode), open-codesign (streaming artifact loop, live preview), and multica (daemon-and-runtime, PATH-scan age). The prompt stack trains the LLM to behave like a senior designer with a working filesystem, deterministic palette library, and checklist culture.

Typical workflow: prompt like "make me a magazine-style pitch deck for our seed round" triggers an interactive question form, the agent picks one of five curated visual directions, streams a TodoWrite plan, builds a project folder, runs pre-flight checks, performs a five-dimensional critique, and emits a single <artifact> rendered in a sandboxed iframe.

Who It's For

Developers who already use AI coding agents (Claude Code, Codex, Cursor, etc.) and want a local, open-source design engine without cloud lock-in or per-seat fees.

📖 Read the full source: HN AI Agents

Ad

👀 See Also

ClaudeClaw: Free Claude Code Plugin for Persistent AI Agents on Messaging Platforms
Tools

ClaudeClaw: Free Claude Code Plugin for Persistent AI Agents on Messaging Platforms

ClaudeClaw is a free, open-source MIT-licensed Claude Code plugin that runs Claude as a persistent agent on Slack, WhatsApp, and Telegram. It requires Node.js 20+, Claude Code, and your own Anthropic API key, with OS-level sandbox isolation via Anthropic's sandbox-runtime.

OpenClawRadar
ShareMyClaudeMD: Tool Converts Claude-Generated Markdown Files to Shareable Rendered Pages
Tools

ShareMyClaudeMD: Tool Converts Claude-Generated Markdown Files to Shareable Rendered Pages

A developer built sharemyclaudemd.com, a free tool that converts any Markdown file into a live, rendered page with a shareable URL and QR code. The tool addresses the friction of sharing Claude-generated Markdown files, which often requires recipients to open them in a specific editor or push to GitHub just for a rendered view.

OpenClawRadar
Multi-provider LLM fallback chain with Ollama support in production AI IDE
Tools

Multi-provider LLM fallback chain with Ollama support in production AI IDE

Resonant Genesis AI IDE integrates local LLM support as a first-class provider alongside Groq, OpenAI, Anthropic, and Gemini across 30+ microservices using a shared UnifiedLLMClient library with automatic fallback chain.

OpenClawRadar
LM Studio plugins add web image analysis for vision-capable LLMs
Tools

LM Studio plugins add web image analysis for vision-capable LLMs

A developer created plugins for LM Studio that enable vision-capable LLMs to fetch and analyze images from the web, with automatic image processing and tool chaining. The plugins work with models like Qwen 3.5 9b/27b and include updated Duck-Duck-Go and Visit Website functionality.

OpenClawRadar