Prompts to Generate a Full SaaS

I have spent the last month testing ALL of the ai coding tools trying to figure out which one is the best and how to best prompt them

and after extensive research, I have found the best prompt, and the best ai coding tool to actually build a fully ready to go software application that you can go out and sell

so as promised, here is my master prompt:

=== BLOCK 1: PROJECT FOUNDATION ===

You are building a complete, production-ready software company application from the following Product Requirements Document (PRD). Your job is to interpret the PRD holistically — not just as a feature list, but as a business with users, workflows, and goals.

[PASTE YOUR FULL PRD HERE]

Before generating anything, confirm your understanding by outputting:

1. The core problem being solved

2. The primary user persona(s)

3. The top 3 critical workflows

4. Any ambiguities or assumptions you are making

Do not start building until this summary is confirmed.

=== BLOCK 2: DATA MODEL & SCHEMA ===

Based on the PRD, design the complete data model for this application. For each entity output:

• Entity name and purpose (1 sentence)

• All fields with types (string, number, boolean, date, enum, relation)

• Required vs optional fields

• Relationships to other entities (one-to-many, many-to-many)

• Any indexes needed for performance

Then generate the database schema (use [PostgreSQL / MongoDB / Supabase — pick one]) reflecting this model. Include:

• Primary keys and foreign keys

• Enums and their allowed values

• Soft-delete fields if applicable

• created_at / updated_at timestamps on all tables

Flag any data modeling decisions that were not explicit in the PRD so I can approve them.

=== BLOCK 3: ARCHITECTURE & TECH STACK ===

Design the full technical architecture for this software company based on the PRD. Output:

FRONTEND

• Framework and rationale

• State management approach

• Key pages/views and their routing structure

• Auth flow (login, signup, password reset, protected routes)

BACKEND

• API design (REST or GraphQL) with the 10 most critical endpoints listed

• Authentication strategy (JWT, sessions, OAuth)

• Background jobs or async processes needed

INTEGRATIONS

• Third-party services required (payments, email, storage, analytics, etc.)

• Webhooks needed

INFRASTRUCTURE

• Hosting recommendation

• CDN and file storage strategy

• Environment structure (dev / staging / prod)

Output this as a structured technical spec, not prose. Flag any decisions the PRD left ambiguous.

=== BLOCK 4: FEATURE BUILD (REPEAT PER FEATURE) ===

Build the following feature from the PRD: [FEATURE NAME]

Context from PRD: [PASTE RELEVANT PRD SECTION]

For this feature, generate:

1. BACKEND — API route(s) with full request/response shapes, validation logic, error handling, and any DB queries

2. FRONTEND — Component(s) with all states: loading, empty, error, success, and edge cases

3. BUSINESS LOGIC — Any calculations, rules, or workflows that live between the API and the UI

4. TESTS — At minimum: one happy path test, one validation failure test, one edge case test

Constraints:

• Follow the data model and architecture defined earlier in this session

• Do not introduce new dependencies without flagging them

• Every form must have field-level validation

• Every API route must have auth checks

After generating, list any open questions or assumptions made.

=== BLOCK 5: AUTH, ROLES & PERMISSIONS ===

Design and implement the full authentication and authorization system for this application based on the PRD.

AUTHENTICATION

• Sign up flow (fields, validation, email verification if needed)

• Login flow (credentials, error messages, rate limiting)

• Password reset flow

• Session management and token refresh

AUTHORIZATION

• List every user role defined in the PRD (e.g. admin, member, viewer, owner)

• For each role, define what they can READ, CREATE, UPDATE, DELETE across each entity

• Output this as a permissions matrix table

MULTI-TENANCY (if applicable)

• How is data scoped per organization/team/account?

• Row-level security rules

IMPLEMENTATION

• Middleware for protecting routes

• Frontend route guards

• How unauthorized access is handled (redirect vs 403 response)

Flag any role or permission not explicitly defined in the PRD — do not invent them without surfacing them first.

=== BLOCK 6: UI/UX & DESIGN SYSTEM ===

Generate a consistent UI design system for this application based on the PRD and the user personas.

DESIGN TOKENS

• Primary, secondary, and accent colors (provide hex values)

• Typography scale (font family, sizes for h1–h4, body, caption, label)

• Spacing scale (4px base grid)

• Border radius and shadow definitions

CORE COMPONENTS — generate each as a reusable component:

• Button (primary, secondary, destructive, disabled states)

• Input field (default, focused, error, disabled)

• Data table (sortable columns, pagination, empty state)

• Modal / dialog

• Toast / notification system

• Navigation sidebar or top bar

• Loading skeletons for async content

• Empty state template

LAYOUT RULES

• Max content width and responsive breakpoints

• Page layout template (sidebar + main, or top nav + main)

• Mobile behavior for key pages

Every component must have all interactive states. Do not generate placeholder/lorem content — use realistic data that matches the PRD's domain.

=== BLOCK 7: ERROR HANDLING & EDGE CASES ===

Audit the application built so far against this checklist and fix any gaps:

API RESILIENCE

• Every endpoint returns consistent error shapes: { error: string, code: string, details?: any }

• 400 for validation errors, 401 for unauth, 403 for forbidden, 404 for not found, 500 for server errors

• No raw database errors exposed to the client

• Timeout handling on external API calls

FRONTEND RESILIENCE

• Every async operation has: loading state, error state, and empty state

• Forms re-enable after submission failure (no stuck loading buttons)

• Network errors show a human-readable message, not a stack trace

• Optimistic updates are rolled back on failure

DATA INTEGRITY

• No orphaned records possible through the UI

• Concurrent edit conflicts handled (last-write-wins or conflict detection)

• Deleted resources return 404, not a blank page

RATE LIMITING & ABUSE PREVENTION

• Auth endpoints are rate limited

• File uploads are size and type validated on the server, not just the client

List every issue found and the fix applied.

=== BLOCK 8: LAUNCH READINESS CHECKLIST ===

Run a final launch readiness review of this application against the original PRD.

PRD COVERAGE

• List every feature or requirement from the PRD

• Mark each as: ✅ Implemented | ⚠️ Partial | ❌ Missing

• For partials and missing items, explain the gap

SECURITY

• Are all routes protected that should be?

• Is user input sanitized everywhere?

• Are secrets stored in environment variables (not hardcoded)?

• Is HTTPS enforced?

PERFORMANCE

• Are database queries using indexes?

• Is pagination implemented on all list endpoints?

• Are images optimized?

OBSERVABILITY

• Is there error logging to a service (Sentry, Datadog, etc.)?

• Are key user actions tracked for analytics?

DOCUMENTATION

• Is there a README with setup instructions?

• Are environment variables documented?

• Is the API documented (even a simple list of routes)?

Output the results as a structured report. Prioritize blockers (must fix before launch) vs improvements (nice to have).

Now it is best to do these section by section, however, you can also one shot it and it still works pretty well

now, which AI coding tool is the best?

I think Base44 is the best, here’s why:

1. It's genuinely all-in-one. Most AI builders make you stitch together a frontend tool, a database service, an auth provider, and a hosting platform separately. Base44 handles the front-end, back-end, and user authentication in a single browser tab

2. Instant deployment, no DevOps. Base44 comes with built-in hosting, so there's no deployment process — when your app is created, it's instantly live and shareable

3. A "discuss before you build" mode. Before committing code, you can chat with the AI architect, adjust logic, or ask "Why did you model the database this way?"

«here is my affiliate link to sign up»

when you sign up with my link and provide proof (just reply with a screenshot or say done) I will hop on a 30 minute call with you once your SaaS is built to go over your go to market strategy

$20/mo plan for 30 min with me, which is usually $250

Go build something this week

Alex