Archify
tt-a1i/archifyReviewed by AI
What it does
Archify is a skill you add to an AI coding assistant such as Claude Code, Codex CLI, Cursor or OpenCode. You describe a system or a plan, or point it at a repository, and the assistant writes a diagram that comes out as one interactive HTML file you can open in a browser, explore, export as an image and send to others. Ten diagram shapes are covered, from architecture and sequence to timelines and database schemas. You need Node.js 18 or newer and one of those assistants. The README says it is not a general-purpose drawing editor, so it does not replace free-form tools like Lucidchart.
How easy it is to install
C · install 1 or 2 tools · Windows
- Windows C
- macOS C
- Linux C
Same work on Windows, macOS and Linux: install Node.js 18 or newer plus a supported AI assistant, then run one command in the terminal (`npx skills add tt-a1i/archify -g`) and ask the assistant for a diagram. There is no installer and no ready binary; the 23 releases carry no platform files.
Numbers on GitHub
- Stars
- 92,154on Oct 31, 2026
- Downloads
- 36,2851,234 this week
- Growth
- +2,341stars in 5 days
Stars over time
Facts
- Runs on
- Windows, macOS, Linux
- You need to install
- Node.js, Other tools
- To build from source
- Node.js
- Disk space
- 100 MB–500 MB · estimate
- Accounts or keys
- None
- Cost
- Free, API costs extra · The project is free under Apache-2.0. It only works inside an AI coding assistant, which normally needs a paid plan or paid API use. The README also links a WeChat Pay and an Alipay QR code for donations.
- License
- Apache-2.0 (Permissive)
- Can replace
- Lucidchart, Visio
- Last update
- Oct 31, 2026
Risk signals (4)
- Attentionreadme sponsor links to model resellers with referral codes
The README Sponsors table promotes several third-party services that resell access to Claude, OpenAI and Gemini models at a fraction of list price, each through a referral link with the project's own code: OpenLux (`openlux.ai/register?channel=c_qdvanbpc`, 'as low as 8.82%, 4% and 8% of official rates'), PackyCode (`packyapi.ai/register?aff=bN97`), AICodeMirror (`aicodemirror.ai/register?invitecode=WC26FL`, 'Codex pricing starts at 7% of standard rates') and Infistar (`infistar.cc/register?aff=9N89LEMV`, 'starts at 1% of official rates'). These are other companies' paid relay services, not part of Archify.
AI review
- Infooptional update check contacts the authors server
README Quick start: 'Archify may GET the fixed stable manifest solely to show an optional reminder; it never downloads or installs updates. (...) The server sees normal HTTP metadata (IP and time), but receives no version, Agent, project data, prompts, account/device ID, or ETag. (...) Set ARCHIFY_UPDATE_CHECK_DISABLED=1 to disable networking and reminder-state writes.'
AI review
- Infozip archive committed in the repository
The top-level file list includes `archify.zip`, which the README tells you to upload under Claude.ai Settings → Capabilities → Skills. The contents of that packaged file are not reviewable as plain text in the repository listing.
AI review
- Infoskill tells the assistant to read your repository
README: 'Analyze this repository, then use archify to create a high-level runtime architecture diagram' and 'Evidence-backed Architecture nodes mark themselves SRC n and open Git-verified files and line ranges pinned to one public commit'. Source-backed diagrams mean your code paths and file excerpts are read by the assistant and can end up inside the HTML you share.
AI review
Evidence is shown in its original language.
What Radar could not check
- Whether the generated diagrams are accurate for a real codebase, since the assistant writes them
- How much AI usage, and therefore cost, one diagram or an explainer page consumes
- Whether the Claude.ai upload path works, since the README says it 'depends on Node.js access in the sandbox'
- What is inside the committed archify.zip compared with the source folders
- Whether the optional update check really sends no project data, as the README states
- The real disk use after installing the skill and its viewer assets