AIAny
AI Agent2026
Icon for item

Diagram Design

Generates editorial-quality diagrams as self-contained HTML files with inline SVG across 27 visual types; includes brand onboarding, agent-skill integrations (Claude Code, Codex, Pi), draw.io/Mermaid import, and static-first output with optional accessible motion.

Introduction

One core workflow problem this project solves: diagrams that match a site's brand and editorial tone instead of generic rounded-box outputs from LLMs. By embedding a small design system into an agent skill, it produces ready-to-use, accessible diagram HTML that reads your site for colors/fonts, redraws draw.io/Mermaid sources to an editorial style, and exports SVG/PNG for slides or Figma without a build step.

What Sets It Apart
  • Inline, offline-first deliverables: every diagram is a self-contained HTML file with inline SVG/CSS (no external images or runtime dependencies), so outputs open and render identically when double-clicked or embedded.
  • Brand-aware onboarding: an automated flow extracts dominant palette and font stack from a website, maps them to semantic tokens (paper, ink, accent), checks contrast, and applies them across all diagram types — reducing manual styling work.
  • Agent-first integration and import tooling: ships as a skill for Claude Code, Codex, and Pi, can redraw draw.io and Mermaid sources at controlled detail/audience/size dials, and emits a fidelity ledger describing what changed in conversion.
  • Accessibility and editorial restraint: accessible SVG naming, reduced-motion defaults, a one-accent editorial rule, and rules for node density and label placement to avoid generic, AI-like visuals.
Who it's for & tradeoffs

Great fit if you embed diagrams into documentation, blog posts, or slide decks and need them to align with your brand or editorial voice — especially teams using agent workflows (Claude/Codex/Pi) who want reproducible, exportable diagram artifacts. It’s also useful when you need faithful-but-simplified redraws of draw.io or Mermaid sources for different audiences (engineer vs executive).

Look elsewhere if you only need quick ASCII/inline diagrams, one-off decorative images, or an interactive/design-heavy Figma-first workflow. The project intentionally avoids being a full-fledged visual editor or replacing dedicated design tools; its strength is reproducible, accessible, editorial-quality static output and scripted agent integration.

Information

  • Websitegithub.com
  • OrganizationsBestSelf.co, Littlemight
  • AuthorsCathryn Lavery
  • Published date2026/04/16

Categories

More Items

Hugging Face

Provides 5.5K+ self-contained data-analysis RL tasks: each row bundles a real tabular dataset, a question, and a deterministically-gradable gold answer. Verified from jupyter-agent notebooks; splits for training, held-out testing, and quick eval; intended for prompting, fine-tuning, and agent RL.

Hugging Face
AI Model2026

A 9B agentic multimodal SFT checkpoint distilled from Qwen3.5-9B for coding, general agent tasks, visual coding and cybersecurity. Provided by Xiaomi MiMo as a research seed (77.4B-token SFT mix) to bootstrap agentic RL and tool-use experiments.

Hugging Face
AI Model2026

Preview agentic language model for research and engineering workflows that turns research questions into executable, verifiable workflows via tool use and long-context reasoning; built on a 744B-parameter MoE (GLM-5.2) with MIT-licensed BF16 and FP8 checkpoints.