What it is
Archify is an open-source skill (MIT) for AI coding agents. You describe an idea, plan, or system, or ask the agent to read a repository, and it generates typed JSON that is validated and rendered as a standalone interactive HTML file. It supports Architecture, Workflow, Sequence, Data Flow, and Lifecycle diagrams, with theme switching, path tracing, and exports.
Who it's for
- Developers who want source-backed architecture diagrams of a repository from their coding agent
- Teams who want to share interactive diagrams as a single HTML file
- Users who want to turn plans such as itineraries or learning maps into explorable visuals
- Developers who want to extend the generated HTML with their own interactions
Requirements
Requirements
- An AI agent such as Cursor, Claude Code, Codex CLI, or OpenCode
- npx, used by the documented install commands
- node, used to run the bundled CLI (node archify/bin/archify.mjs)
- A network connection only for external links in artifacts and the optional update-reminder check
Setup
Install the skill
Install Archify globally for your agent.
bashnpx skills add tt-a1i/archify -gExplicit non-interactive Cursor install
Use this for a scripted install targeting Cursor.
bashnpx -y skills add tt-a1i/archify --skill archify --agent cursor --global --copy --yesTry without installing
Run the skill for Codex without a permanent install.
bashnpx skills use tt-a1i/archify@archify --agent codexOptionally disable the update check
Archify may fetch a stable manifest to show an optional reminder. Set the ARCHIFY_UPDATE_CHECK_DISABLED=1 environment variable to disable networking and reminder-state writes.
Examples
Diagram from a description
PromptUse Archify to diagram a web request: Browser calls the API,
the API checks Redis, and a cache miss queries PostgreSQL and fills the cache.Expected output: No repository is needed. The agent generates the interactive diagram from the description.
Source-backed repository diagram
PromptAnalyze this repository, then use archify to create a high-level runtime architecture diagram.
Show 8–12 core components, one primary path, external dependencies, and trust boundaries.
Put supporting detail in cards instead of adding more edges.Expected output: Run this inside a repository to get a runtime architecture map.
Pick a diagram type with the CLI guide
bashnode archify/bin/archify.mjs guide "Show an API request with Redis cache miss"What it does: The zero-dependency CLI suggests which diagram type fits a scenario.
Validate and deliver a diagram
bashnode bin/archify.mjs validate workflow examples/agent-tool-call.workflow.json --quality showcase --json
node bin/archify.mjs deliver workflow examples/agent-tool-call.workflow.json /tmp/workflow.html --quality showcase --open --jsonWhat it does: Run from the archify directory. Validate checks the typed source, and deliver renders and checks the HTML before atomically replacing the target.
Refine in chat
PromptAdd authenticationExpected output: Follow-up requests like this, or "Highlight the cache-miss path" and "Switch to the light theme", adjust the existing diagram.
Pros & cons
Pros
- Pro:Output is one self-contained HTML file that can be opened or shared without installing Archify
- Pro:Validation gates (schema, layout, HTML/SVG, routes) must pass before a showcase artifact replaces the last good output
- Pro:Five diagram types plus Architecture Delta comparison, route/reach exploration, and PNG/share-card exports
- Pro:Free, MIT-licensed, and works with Cursor, Claude Code, Codex CLI, and OpenCode
Cons
- Con:Not a general-purpose drawing editor or a Mermaid theme; it produces communication artifacts from technical intent
- Con:Locales other than en and zh-CN need meta.translations, otherwise it falls back to English
- Con:Updates are never installed automatically; users must update manually, and older installs without the checker get no reminders
Images
Related
Repo DietrichGebert/ponytail
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Repo stablyai/orca
Orca is the ADE for working with a fleet of parallel agents. Run any coding agent with your own subscription. Available on desktop, mobile and remote runtime.
Repo trailhq/Graft
Turbocharge Claude Code, Cursor, Codex, Gemini & every coding agent: faster, cheaper, with contextual understanding specific to your codebase.
Repo affaan-m/ECC
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
Repo farion1231/cc-switch
A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io
Repo Graphify-Labs/graphify
Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local deterministic AST parsing, every edge explained, no vector store.