Makimono

Don't build from scratch.
Feed your AI the scroll.

Building from zero burns ~100K tokens (~$1). With the right scroll, your AI reads it and ships for ~3K (~$0.03).

Makimono is a marketplace for AI instruction files — CLAUDE.md, skills, and dev playbooks. Stop paying for your AI's trial-and-error.

Use every scroll on the shelf and save4.7Mtokens$47 in API cost · 19 playbooks and growing

💡 Tokens = what you pay your AI. The more it reads and retries, the more it costs.

What's a “scroll”?

A finished recipe: the instructions, tech choices, and gotcha-avoidance that make an AI build the thing on the first try.

1. Find

Search by what you want to build — or let your agent find it automatically through the open API.

2. Feed it in

Download the .md and hand it to Claude Code, Cursor, or any agent. One paste, no setup.

3. Ship

Your AI heads straight to done instead of wandering. The tokens (= money) it would burn on trial-and-error are saved.

How is this different from a skills marketplace?

A skill (SKILL.md) is a narrow, packaged capability — it teaches an agent how to do one specific thing. A scroll is broader: it's the accumulated judgment behind an entire project — the tech choices, the gotchas, the “actually, do it this way” calls that only show up after you've been burned once. Skills are welcome here too (some listings are skills), but what Makimono is really built to collect is judgment, not just capability.

Zero busywork. AI does the listing and the buying.

You barely touch it. Your AI finds, reads, implements, and even lists — automatically.

Listing = paste one prompt

Got a CLAUDE.md that just works? Paste one prompt into your AI and it collects your best instruction files, strips secrets (API keys, client names), drafts the description, and submits — all with a per-item approval gate. Zero manual work. Earn when it sells.

Get the listing prompt →

Buying is automatic too

Set it up once and every task, your AI finds the best scroll, loads it, and builds. You just say “build X.” No human has to browse. Runs on the open, no-auth API.

See the API →

The more it's used, the more it grows — a flywheel, not a pile of files

Listings (supply) and usage (proven savings) reinforce each other. Every turn makes the next turn easier.

① List

A scroll goes on the shelf

② Use

An agent reads it, savings get logged as proof

③ Trust

Proof builds up — contributors want in (and earn)

④ Grow

More scrolls match more tasks → back to ①

A scroll collapses your agent's trial-and-error loop into one pass. Multiply that across contributors and it becomes a flywheel.

Your AI can use this site by itself

An open, no-auth, CORS-enabled API lets any agent find the right scroll and load it — no human has to search per task.

# Find a scroll for what you're building
curl "https://makimono-md.vercel.app/api/v1/search?q=discord+bot"

# → read results[0].links.raw, then start building
curl "https://makimono-md.vercel.app/api/v1/files/\
{slug}/raw"

Launching from Japan 🇯🇵 — help us go global

Many current listings are written in Japanese. English CLAUDE.md files are especially welcome as we expand — be one of the first scrolls on the global shelf.