AI-powered product planning for builders

Your next product, fully planned before you write a line of code

Get a complete technical blueprint — database schemas, API endpoints, architecture decisions, and per-feature specs — generated from a conversation or your existing GitHub repo.

shipblueprint.com/blueprint/my-saas-app

My SaaS AppPlan

v1·Feb 25, 2026

Files

docs

MPLAN.md

MARCHITECTURE.md

MDATA-MODEL.md

MAPI.md

MBUILD-ORDER.md

MFILE-STRUCTURE.md

features

Muser-auth.md

Mdashboard.md

Mbilling.md

docs/PLAN.mdCopy

# My SaaS App

A project management tool for small teams that combines task tracking with real-time collaboration.

## Problem Statement

Small teams (3-10 people) struggle with existing PM tools that are either too complex (Jira) or too simple (sticky notes). They need something in between...

## Value Proposition

Lightweight project management with built-in team chat. No switching between apps — discussions happen right on the task.

## Target Users

1. Startup Team Lead — Manages 5-8 engineers, needs sprint planning without overhead...

Not a summary. A full technical spec.

Every file is dense, actionable, and specific to your project. Here's what's inside each one.

docs/DATA-MODEL.mdEvery table, every column

## users

Column	Type	Nullable	Description
id	uuid	no	PK, default gen_random_uuid()
email	text	no	Unique, from Clerk
org_id	uuid	yes	FK → organizations.id
role	enum	no	admin \| member \| viewer

+ tasks, projects, comments, attachments, billing...

docs/API.mdEvery endpoint, typed

## Tasks

GET/api/tasks— List tasks for project

Response: { tasks: Task[], total: number, cursor?: string }

POST/api/tasks— Create task

Body: { title: string, assignee_id?: string, priority: 1-4 }

PATCH/api/tasks/:id— Update task

docs/features/team-chat.mdOne file per feature — user stories, UI, acceptance criteria

# Team Chat

As a team member, I want to discuss tasks in real-time without leaving the app, so that context stays with the work.

## How It Works

Each project has a general channel. Each task has a thread. Messages support @mentions (notify via Supabase Realtime) and file attachments (upload to S3, store reference in messages table).

## UI Components

ChatSidebar, MessageList, MessageInput, ThreadView, MentionPopover, AttachmentPreview

## Acceptance Criteria

✓ Messages appear in under 200ms for all connected users

✓ @mention triggers browser notification if recipient is offline

✓ Thread replies don't clutter the main channel

✓ File uploads limited to 10MB, preview for images

docs/BUILD-ORDER.mdPhased plan with dependencies and sizing

## Phase 1 — Foundation

SProject setup, Next.js + Tailwind + Supabase

SClerk auth with org support

MDatabase schema + RLS policies

MBase layout, nav, dashboard shell

## Phase 2 — Core MVP

LTask CRUD with drag-and-drop board

MReal-time team chat (WebSocket)

SProject settings + invite flow

docs/PLAN.mdVision, audience, scope — the why behind the what

## Problem Statement

Small teams (3-10 people) bounce between Jira, Slack, and Google Docs. Context gets lost. Decisions live in chat threads nobody can find.

## Target Users

1. Startup Tech Lead — Manages 5-8 engineers, runs weekly sprints, needs lightweight planning without the overhead of enterprise tools.

2. Freelance Team — 3-4 contractors on a project, need shared visibility into who's doing what without onboarding to complex software.

## Scope Boundaries (v1)

✗ No Gantt charts, no time tracking, no resource allocation. ✗ No mobile app — responsive web only.

Two ways to get your blueprint

Drop the docs/ folder into your project root. Point Claude Code, Cursor, or Copilot at it. Start building with full context.

An AI coach guides you through 5 phases — vision, audience, features, architecture, and generation. Answer questions, pick from options, and watch a visual canvas map your product in real-time.

Guided conversation

Choose Quick (5 min), Thorough (15 min), or Adaptive mode. Clickable options at every step — no blank-page anxiety.

Visual product canvas

An interactive diagram builds in real-time as features, audience, and architecture decisions emerge from the conversation.

Per-feature specs

Every feature gets its own file — user stories, UI components, API calls, and acceptance criteria. Your AI loads only the context it needs.

Phased build order

Not a flat list — a phased plan with dependencies and S/M/L sizing. Foundation first, then core MVP, then polish.

Typed API reference

Every endpoint with method, path, request/response shapes, and validation rules. TypeScript-style types your AI turns into real code.

Complete data model

Every table, every column, every foreign key. Formatted as markdown tables so AI tools can generate migrations directly.

PlanOutput

docs/