Skip to content

Aprasaks/dechive

Repository files navigation

Dechive

AI는 답을 만든다. Dechive는 그 답을 검증한다.

dechive.dev


What is Dechive?

Dechive is a personal archive for turning AI-era answers into verifiable records.

It is not a general blog, portfolio, or tutorial platform. Archive keeps one question as an independent record. Deep Dive carries a deeper question through concepts, examples, mistakes, limits, reasoning, and verification criteria. Downloads turns selected criteria into reusable material.

In the age of AI, information is infinite and cheap. Verified knowledge, written from first principles and checked against reality, is not.


Content Model

type       — archive | deepdive
category   — broad content area
post       — independent record or long-form verification document
tags       — index terms
concepts   — concept links for structured reasoning

Current content types:

  • Archive — independent records built around one question
  • Deep Dive — long-form documents built around one deep question
  • Downloads — reusable ebooks and materials

Content Pipeline

Content lives in a separate Git submodule, decoupled from the application code.

# Write the Korean post
content/posts/{slug}.ko.md

# Generate the English translation
npm run translate
# -> content/posts/{slug}.en.md

# Commit and push content first
cd content && git add . && git commit -m "feat: ..." && git push

# Then update the root submodule pointer
cd .. && git add content && git commit -m "chore: update content submodule" && git push

Tech Stack

Layer Technology Why
Framework Next.js App Router SSG + SEO control
Language TypeScript Strict application contracts
Styling Tailwind CSS v4 Responsive interface work
Guestbook Supabase Lightweight visitor notes
Translation Anthropic API KO -> EN post translation
Content Git Submodule + Markdown Code and content decoupled
Deployment Vercel Production hosting

Principles

  • Verification first — Every record exists to make an answer checkable.
  • Independent records — Archive posts must stand on their own.
  • Deep reasoning — Deep Dive documents carry one question through limits and criteria.
  • Clear boundaries — Automation should support publishing, not obscure it.

Dechive

About

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors