Archscribe logo

Archscribe

Community
lazypay
archscribe

Create premium hand-drawn architecture, workflow, and swimlane diagrams in a dark neon or light paper style, with editable Excalidraw sources, PNG/SVG previews, animated GIF/MP4 output, and interactive HTML. Use whenever the user asks for 岚叔动态架构图、手绘架构图、Excalidraw 风格图、DailyDoseOfDS 风格技术图、动态流程图、泳道图、系统可视化,或希望复刻并提升参考架构图。

Overview

Publisherlazypay
RepositoryArchscribe
Skill namearchscribe
Stars
369
Forks
24
Bundled files
78
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 78 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by lazypay on GitHub. Read the source before you install it.

Installation

Install the Archscribe AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/lazypay/Archscribe/tree/main
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/lazypay/Archscribe.git \
  .claude/skills/archscribe
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Archscribe in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Archscribe on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Archscribe is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Archscribe

Turn an article, system description, process, or reference image into a polished hand-drawn diagram. The default browser renderer can produce PNG, GIF, MP4, SVG, editable Excalidraw, and interactive HTML.

1. Check the environment

Run once per session:

bash
python -X utf8 -c "import PIL, svg.path"
python -X utf8 -c "from playwright.sync_api import sync_playwright"
python -X utf8 scripts/doctor.py
  • Missing base packages: python -m pip install -r requirements.txt
  • Missing Playwright: python -m pip install -r requirements-browser.txt && python -m playwright install chromium
  • Missing ffmpeg: MP4 is skipped; other browser formats still work.
  • Chromium unavailable: use --renderer pillow. Pillow supports PNG, GIF, and Excalidraw with the classic flow animation.

Bundled fonts, icons, and rough.js keep the main visual system portable. Chromium, ffmpeg, and rare CJK fallback fonts remain environment dependencies.

2. Pick a layout

LayoutUse it forCapacity
panoramacomplete systems with inputs, core, decisions, and detail panels2-6 inputs, 2-4 core cards, 0-3 panels
swimlanecategory bands / comparison rows (DailyDoseOfDS-style), cross-role workflows, catalogs2-5 lanes with title + subtitle column, up to 5 steps each, in-lane dashed loop channels
graphfree-form per-project workflows with custom loop topology (retry/replan/recall...)2-24 nodes, up to 40 edges, auto DAG layout + loop lanes

Decision rule:

  • Whole system with multiple regions → panorama
  • Parallel categories, variants, roles, or a comparison ("the N types of X") → swimlane (each lane gets a tinted band, a darker title column with optional subtitle such as "Triggered by: ...", and right-to-left in-lane connections automatically drop into a dashed loop channel under the cards)
  • Anything with its own topology: linear stages, forks/joins, multiple distinct loops → graph (declare nodes + edges; kind: "loop" edges become dashed return channels that fire after the forward wave). Long rightward graphs auto-stack downward to prevent clipped sides and top-heavy layouts.

3. Write and validate the spec

Start from assets/default-spec.json or a file under assets/examples/. The complete field reference is references/spec-format.md.

Keep titles short, use the user's language, and move detail into bodies, notes, or an interactive HTML sidebar.

bash
python -X utf8 scripts/render_animated_diagram.py --spec spec.json --outdir out --validate-only

Fix every error. Warnings identify long text, unknown icons, ignored fields, content that exceeds a layout's capacity, graph auto-stacking, or ignored graph canvas width/height.

4. Choose a visual theme

Two styles, both first-class:

  • default: black-canvas neon hand drawing — glowing beams, grain, vignette. The brand hero look.
  • paper: light warm-white paper (DailyDoseOfDS-style) — sage/periwinkle band tints, near-black ink, white cards with colored strokes; the flow animation becomes small solid dots traveling the arrows instead of neon beams.

Styles control the full palette and finish. Use --style to override the spec. Any other style name fails validation.

5. Choose animation

  • flow: continuous data movement; safe default
  • draw: whiteboard construction
  • relay: one connection at a time
  • trace: follow a request path
  • chapter: staged explanatory build-up
  • failure-recovery: failure/retry-oriented narrative timing

The last three use deterministic narrative choreography and longer minimum timelines. The browser renderer is required for every preset except classic Pillow flow.

6. Render and verify

First render a PNG for layout review:

bash
python -X utf8 scripts/render_animated_diagram.py --spec spec.json --outdir out --basename diagram --formats png --check

Then render the publishing set:

bash
python -X utf8 scripts/render_animated_diagram.py --spec spec.json --outdir out --basename diagram --formats gif,mp4,png,excalidraw,svg,html --verify --check --strict-formats

--formats is exact: only requested, supported files are emitted. Unknown formats fail. Use --strict-formats for publishing so missing MP4/SVG/HTML fails instead of silently falling back. --check must return "ok": true.

Visually inspect the PNG for overlap, crossing connectors, cramped labels, weak hierarchy, clipped edges, and empty-space imbalance. Shorten copy or switch layout before reducing text below a readable size.

7. Icons and brands

Specs use stable semantic icon keys. Core keys include file, folder, scan, shield, db, package, message, api, clock, brain, gear, eye, terminal, globe, server, lock, check, and clipboard.

Additional semantic aliases cover cloud infrastructure, data, AI, security, business roles, and states, including cloud, cluster, container, queue, cache, vector-db, agent, model, rag, tool-call, guardrail, evaluation, identity, audit, user, success, failure, and retry.

For a custom logo or colorful icon, use icon_file with a local SVG/PNG path. Use left_panel.badge_file for a panorama panel brand mark. Paths resolve relative to the spec.

Illustration system

Every item that accepts icon also accepts:

  • icon_style: outline, illustrated, or hero.
  • icon_size: compact, standard, or hero.
  • icon_motion: auto (plays the icon's built-in job story) or none (freeze at rest pose). Legacy preset names (think-pulse, gear-spin, eye-scan, memory-write, shield-check, scope-scan, budget-gauge, trigger-ping, tool-spark, output-pop) are still accepted and behave like auto.

Use outline for dense secondary nodes, illustrated for normal concept cards, and hero for the 1-3 concepts that carry the story. The browser renderer draws plate-free "Neon Sketch Duotone" illustrations: theme-ink hand-drawn strokes plus one semantic accent on the moving part, where each of ~56 semantic families performs its own seamless-looping job story (synapse signals, gear pitches, chip drops, shackle clicks — see references/illustrated-icons.md). This includes a dedicated loop-workflow pack (loop, decision, split, merge, wait, orchestrator, subagent, handoff, human, plan, score, compare, sandbox, checkpoint, error, rollback, emit, ingest, …) for agent-loop and pipeline diagrams — sample spec assets/examples/loop-icon-pack-spec.json. Pillow renders a simplified offline duotone fallback; Excalidraw keeps editable icon placeholders.

For a DailyDoseOfDS-style short loop, use 1210×1138, 20 FPS, 41 frames, stable camera, 1-3 active paths at a time, illustrated/hero icons on the main concepts, and flow animation. Start from assets/examples/illustrated-loop-spec.json (dark panorama) or assets/examples/swimlane-spec.json (light paper bands).

8. Delivery

  • Show GIF inline when useful.
  • Prefer MP4 for publishing and social platforms.
  • Include PNG for quick viewing.
  • Include .excalidraw when editability matters.
  • Offer interactive HTML for exploration or presentations.
  • Use SVG for scalable static publication.

After changing this SKILL.md, open a new task or refresh the session before testing trigger behavior. Script and asset changes can be tested immediately.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

and 18 more files.

Frequently asked questions

What does the Archscribe AI skill do?

Create premium hand-drawn architecture, workflow, and swimlane diagrams in a dark neon or light paper style, with editable Excalidraw sources, PNG/SVG previews, animated GIF/MP4 output, and interactive HTML. Use whenever the user asks for 岚叔动态架构图、手绘架构图、Excalidraw 风格图、DailyDoseOfDS 风格技术图、动态流程图、泳道图、系统可视化,或希望复刻并提升参考架构图。

Why use Archscribe on TypingMind?

Because you install it once and use it with any model. Archscribe is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Archscribe in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/lazypay/Archscribe/tree/main. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Archscribe?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Archscribe?

As many as you like. As long as a model supports skills, you can use Archscribe with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Archscribe AI skill free?

Yes. It is published on GitHub by lazypay under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇