AI workflows & creator economy

AI Prompt Library

A versioned prompt library where teams organize, fork, and test prompts across multiple models, then share or sell the ones that work.

Full planning bundle — BRAND, MRD, BRD, PRD, DESIGN
Branded styleguide — typography, color, components
5 previewable mockups

Cloning copies the full plan into a new project you own — edit it freely.

Inside the bundle

Styleguide and mockups

A look at the brand palette and screen-level mockups that ship with this plan. Cloning copies all of them into your project.

Mockups

Style guide

Styleguide

What this blueprint gives you

The “prompt management” model done seriously: every prompt is a versioned object with variables, test cases, and side-by-side runs across model providers. Teams move prompts out of scattered Notion docs and Slack pins into a single library with ownership, history, and review — and creators can publish public packs for a community or sell them.

Five planning documents (BRAND, MRD, BRD, PRD, DESIGN) at senior-team depth, covering the prompt object, multi-model evaluation, forking and sharing, marketplace publishing, and per-seat or per-pack monetization.

Group

AI tools

Preview the plan

Each blueprint includes the senior-level planning files. These cards show the opening text before you clone the full bundle.

Brand Guidelines Document

BRAND.md

Preview

BRAND — AI Prompt Library One line Prompts as source code: versioned, testable, owned by the team that wrote them. Who it speaks to Product, marketing, ops, and engineering teams who already depend on LLMs daily and have realized their best prompts are buried in chat threads, Notion pages, and one teammate's head. Also creators selling prompt packs who are tired of running their store as a PDF folder. Voice & tone Engineering credible. We talk diffs, evals, providers, and tokens — not "AI magic"

Design Specification Document

DESIGN.md

Preview

DESIGN — AI Prompt Library Information architecture Public: marketing page, marketplace listing browse ( /m ), creator profile pages, pack detail pages with previews. App shell (workspace): Library (prompts tree + search), Prompt detail (editor + run drawer + cases), Evaluations (history + trends), Packs (drafts + published), Marketplace (browse + purchases), Settings (members, keys, billing, budgets). Creator studio: within the workspace — pack composer, listing editor, payouts, takedown center

Market Requirements Document

MRD.md

Preview

MRD — AI Prompt Library The problem Teams now write hundreds of prompts a quarter, but treat them like throwaway snippets. The good ones live in a single person's Notion. Comparing models means pasting the same text into three tabs. When a prompt regresses after a model update, nobody can say which version was working last week. The result is duplicated work, unreliable output, and zero institutional memory of how the team actually uses LLMs. Who it's for Primary — the operator/team lead. Heads

Business Requirements Document

BRD.md

Preview

BRD — AI Prompt Library Business outcomes 1. Per seat workspace revenue. Paid seats on team workspaces are the durable line; everything else is a funnel into it. 2. Marketplace take rate. Public pack sales (creator earns, platform takes a percentage) — high margin and viral. 3. Provider neutrality as a moat. Multi model comparison is the wedge against single provider tools; depth of supported providers is a competitive metric. 4. Library size per workspace. Number of versioned prompts per seat —

Product Requirements Document

PRD.md

Preview

PRD — AI Prompt Library Phase 0 — Foundations Auth (email + magic link + Google), workspaces, seats, roles ( owner , editor , viewer ). Provider key vault: add OpenAI / Anthropic / a generic OpenAI compatible endpoint; store encrypted; never exposed in UI after save. Workspace settings: budget, default model, default parameters, retention window for run logs. Acceptance: an owner creates a workspace, invites a teammate as editor, adds a provider key, and the connection test succeeds without expo