LIVE Β· 5 LAYERS Β· ALL PROMPTS Β· BY SHUBHAM
From Any Idea to
Stable Business β€” The OS
A complete orchestration system: 5 ecosystem layers, 100+ prompts, alternate variants, tool stack, and client loop β€” all designed to be run with Claude Opus/Sonnet as your AI co-founder.
The 5-Layer Ecosystem
Each layer is independent but feeds the next. Build in order.
1

Core AI Engine

Idea validation Β· Market research Β· Problem-solution fit Β· Persona mapping

START HERE
2

Product Planning

PRDs Β· Roadmaps Β· User stories Β· MVP scoping Β· Feature prioritization

WEEK 1–2
3

Tech Architecture

Stack decisions Β· System design Β· DB schema Β· API design Β· Infrastructure

WEEK 2–3
4

Business Model + GTM

Pricing Β· Revenue streams Β· GTM strategy Β· Unit economics Β· Competitors

WEEK 3–4
5

Build + Launch Engine

Code generation Β· Testing Β· Deploy Β· Marketing copy Β· Community bootstrap

WEEK 4+

Master Orchestration Flow
The exact order to run everything. Do not skip steps.
Phase 1 β€” Discover
01

Run the Idea Extraction Prompt

Get the raw idea out of the client's head in structured form

02

Market Validation Prompt

Claude searches for market size, existing solutions, gaps

03

Persona Generation Prompt

Define 2–3 ICPs with jobs-to-be-done

04

Feasibility Scorecard Prompt

Score: market, tech, capital, timing, competition

Phase 2 β€” Plan
05

PRD Generation Prompt

Full product requirements document in one shot

06

MVP Scope Prompt

Define the smallest thing that proves the thesis

07

Tech Stack Decision Prompt

Recommend stack based on constraints and timeline

08

Business Model Prompt

Define monetization, pricing tiers, unit economics

Phase 3 β€” Build
09

Architecture Design Prompt

System diagram, data flow, API contracts

10

Component Build Prompts

Auth, DB, frontend, backend β€” module by module

11

Integration Prompt

Connect all modules, third-party APIs, payments

12

QA + Test Prompt

Edge cases, error handling, security review

Phase 4 β€” Launch
13

Launch Copy Prompt

Landing page, Product Hunt post, cold email

14

Community Bootstrap Prompt

Discord/Reddit/X strategy for first 100 users

15

Feedback Loop Prompt

User interview template, survey, NPS flow

16

Iteration Sprint Prompt

Weekly sprint planning from user feedback


The Golden Rule of Orchestration
Every prompt should output structured data (JSON, tables, or markdown) that becomes the input context for the next prompt. Never start a new session cold β€” always pass forward the output of the previous stage. Build a running product document that grows with each step. This document is your product's brain, and Claude reads it at the start of every session.
Layer 1 β€” Idea β†’ Validation
Extract, stress-test, and validate any idea before writing a single line of code.
PROMPT V-01
Idea Extraction & Structuring
Use first. Pulls a structured idea out of vague client thinking.
CRITICAL
You are a world-class product strategist. I have a business idea I want to explore. Help me structure it by asking me 10 sharp questions β€” one at a time β€” that will uncover: 1. The exact problem being solved 2. Who experiences this problem most acutely 3. How they currently solve it (workarounds) 4. What makes this idea unique 5. The scale of the opportunity After my answers, output a structured IDEA BRIEF in this format: - Problem Statement (1 sentence) - Target User (1 sentence) - Proposed Solution (2 sentences) - Unique Angle (1 sentence) - Biggest Risk (1 sentence) - Opportunity Size Guess (rough) My idea: [PASTE YOUR RAW IDEA HERE]
β–Ά Show 2 alternate versions
Alt A β€” Speed version (single shot, no Q&A)
You are a startup validator. Given this raw idea: [IDEA] Output a structured brief with no back-and-forth: - Core problem, target user, solution, moat, risk, opportunity size. Be direct, use bullet points, no fluff.
Alt B β€” Devil's advocate version
You are a VC who has seen 10,000 pitches and rejects 99% of them. I'm going to pitch you this idea: [IDEA] Tear it apart. Tell me every reason it will fail. Then β€” and only then β€” tell me the 1–2 things that could make it work anyway. Be brutal.
PROMPT V-02
Market Size & Opportunity Analysis
TAM/SAM/SOM analysis with web search enabled.
USE WEB SEARCH
You are a market research analyst. Use web search to research the market for: Product Category: [CATEGORY] Target Geography: [GEOGRAPHY] Target User: [USER TYPE] Find and calculate: 1. TAM β€” Total Addressable Market (cite sources) 2. SAM β€” Serviceable Addressable Market (realistic subset) 3. SOM β€” Serviceable Obtainable Market (year 1–3 target) 4. Market growth rate (CAGR) 5. Top 5 existing players and their funding/revenue 6. Market gaps that are underserved Output as a clean table + 3-sentence market narrative.
β–Ά Show 2 alternate versions
Alt A β€” Niche market version
Research a niche market for [IDEA]. Focus on: adjacent markets, underserved segments, and unconventional competitors (people solving this differently). Find the 'hair on fire' segment β€” users who need this most urgently.
Alt B β€” Trend analysis version
Research macro trends driving demand for [IDEA]. Use web search to find: Google Trends data, recent news, VC investment trends in this category, and regulatory tailwinds/headwinds. Output: Is now the right time? Why or why not?
PROMPT V-03
Competitor Deep-Dive
Map the competitive landscape and find the whitespace.
USE WEB SEARCH
You are a competitive intelligence expert. Research competitors for: Idea: [IDEA] Find the top 8 competitors (direct + indirect). For each, output: | Competitor | Founded | Funding | Pricing | Target User | Key Feature | Weakness | Then: 1. Draw a 2x2 positioning map (Price vs Feature Depth) 2. Identify the positioning gap where our product can win 3. List 3 features competitors all have (table stakes) 4. List 3 features no competitor offers well (our opportunity) End with: "Our defensible position should be..."
β–Ά Show alt version
Alt A β€” SWOT for each competitor
For the top 3 competitors to [IDEA], do a SWOT analysis for each. Then answer: If we copied their strength and fixed their weakness β€” what product would that be? That product is our v1.
PROMPT V-04
User Persona Generation
Create 3 detailed ICPs with jobs, pains, and willingness to pay.
HIGH VALUE
You are a user research expert. Create 3 detailed Ideal Customer Profiles for: Product: [PRODUCT] Problem solved: [PROBLEM] For each persona, define: - Name, Age, Job Title, Company Size - Daily workflow (relevant parts) - Top 3 frustrations related to our problem - How they currently solve it (tools, workarounds) - What a 10x better solution would mean for them - Willingness to pay (monthly, annual) - Where they hang out online (channels to reach them) - What they'd Google to find us - Emotional buying trigger ("I'd pay for this the day it could...") Format each as a persona card. Then rank: which persona should we build for first and why?
β–Ά Show alt version
Alt A β€” Anti-persona (who NOT to build for)
Create 2 anti-personas for [PRODUCT] β€” users who seem like our market but will churn, demand too much support, or never pay. Describe who they are and why we should explicitly exclude them from our ICP.
PROMPT V-05
Feasibility Scorecard
Score the idea across 8 dimensions before committing capital.
GATE CHECK
Score this business idea across 8 dimensions. Use the context below. Idea: [IDEA BRIEF FROM V-01] Market research: [PASTE V-02 OUTPUT] Founders: [SKILLS / TEAM] Capital available: [BUDGET] Timeline goal: [TIMELINE] Score each 1–10 with a 1-line rationale: 1. Market Size (is there enough demand?) 2. Problem Urgency (do people need this now?) 3. Technical Feasibility (can it be built with available resources?) 4. Capital Efficiency (can we validate cheaply?) 5. Competitive Moat (can we defend against copycats?) 6. Team-Market Fit (are we the right people for this?) 7. Revenue Clarity (is the path to $1 obvious?) 8. Timing (is the market ready?) Total score: /80. Verdict: GO / NO-GO / PIVOT with reasoning.
Layer 2 β€” Product Planning
Turn a validated idea into a precise product spec that builders can execute on.
PROMPT P-01
Full PRD Generation
One prompt β†’ complete Product Requirements Document.
CRITICAL
You are a senior Product Manager at a top-tier startup. Using this context: - Idea Brief: [V-01 OUTPUT] - Primary Persona: [V-04 PERSONA 1] - Feasibility Score: [V-05 OUTPUT] Write a complete PRD with these sections: ## 1. Product Vision One sentence that defines the north star. ## 2. Problem Statement The exact problem, who has it, and what it costs them today. ## 3. Goals & Success Metrics 3 business goals + KPIs for each (measurable, time-bound). ## 4. User Stories Top 10 user stories in format: "As a [persona], I want to [action] so that [outcome]." Include acceptance criteria for each. ## 5. Feature List Organize as: Must Have / Should Have / Nice to Have (MoSCoW). For each feature: name, description, user benefit, complexity estimate (S/M/L). ## 6. Out of Scope 5 things we explicitly will NOT build in v1. ## 7. Open Questions Top 5 things we need to validate with users before building. ## 8. Launch Criteria What must be true for us to call this MVP "launched"?
β–Ά Show alternate versions
Alt A β€” One-page lightweight PRD (for speed)
Write a 1-page PRD for [IDEA]. Max 500 words. Include only: problem, users, top 5 features, success metric, and what we're NOT building. This should fit on a single Notion page.
Alt B β€” Jobs-to-be-done framework PRD
Write a JTBD-first PRD for [IDEA]. Structure every feature around: "When [situation], users want to [job], so they [outcome]." Do not mention technology. Focus purely on the job the product is hired to do.
PROMPT P-02
MVP Scope Definition
Cut ruthlessly. Define the smallest thing that proves value.
ANTI-BLOAT
You are a lean startup expert. Your job is to REDUCE scope, not add to it. Given this PRD: [P-01 OUTPUT] Answer these questions ruthlessly: 1. If we could only ship ONE feature β€” which one delivers the core value promise? 2. For each Must-Have feature: what is the 80% version that takes 20% of the time? 3. What assumptions are we making that we haven't tested yet? 4. What is the riskiest assumption? How do we test it in 1 week without building anything? 5. Draw a "walking skeleton" β€” the thinnest possible end-to-end flow that a real user could use. Output: MVP Feature Set (max 5 features), Timeline estimate, and ONE metric that proves the MVP worked.
PROMPT P-03
Roadmap Generation (3 Phases)
Now / Next / Later roadmap with rationale.
STRATEGIC
Create a 3-phase product roadmap for [PRODUCT]. Context: [PRD + MVP SCOPE] Team size: [TEAM] Timeline: [MONTHS] Phase 1 β€” NOW (Month 1–2): MVP. Survive. Phase 2 β€” NEXT (Month 3–5): Retention + first revenue. Phase 3 β€” LATER (Month 6–12): Growth + moat. For each phase: - Features shipped - Key metric to move - What we learn - Resource requirement - Go/No-go criteria to advance to next phase Format as a table. Add a risk column for each phase.
PROMPT P-04
User Story Generator
Generate complete user stories with acceptance criteria for any feature.
Generate complete user stories for this feature: Feature: [FEATURE NAME] Product: [PRODUCT] Primary User: [PERSONA] For each story: - User story: "As a [role], I want [action] so that [outcome]" - Acceptance criteria (Given / When / Then format) - Edge cases to handle - Out of scope for this story - UI/UX considerations - API needed (if any) - Estimated complexity: XS / S / M / L / XL Generate: 1 happy path story + 2 edge case stories + 1 error state story.
PROMPT P-05
Feature Prioritization (RICE Score)
Objectively rank your backlog using RICE framework.
Score and rank this feature backlog using the RICE framework. Product context: [PRODUCT + USERS] Features to score: [LIST ALL FEATURES] For each feature, estimate: - Reach: How many users affected per quarter? (number) - Impact: How much does it move our core metric? (0.25 / 0.5 / 1 / 2 / 3) - Confidence: How sure are we? (%) - Effort: Person-weeks to build RICE Score = (Reach Γ— Impact Γ— Confidence) / Effort Output as a ranked table. Flag the top 3 as "Sprint 1" and explain why they're the right bet.
Layer 3 β€” Tech Architecture
Make smart stack decisions and design systems that scale without over-engineering.
PROMPT T-01
Tech Stack Recommendation
Get a justified, constraint-aware stack decision.
CRITICAL
You are a senior CTO advisor. Recommend a tech stack for: Product type: [WEB APP / MOBILE / API / SAAS] Expected users (year 1): [NUMBER] Team skills: [LANGUAGES / FRAMEWORKS KNOWN] Budget: [MONTHLY INFRA BUDGET] Timeline to MVP: [WEEKS] Core features: [TOP 5 FEATURES] Recommend: 1. Frontend framework + why 2. Backend framework + why 3. Database (primary + cache) + why 4. Auth solution + why 5. File storage + why 6. Email service + why 7. Payments + why 8. Hosting/deploy + why 9. Monitoring + why 10. AI/LLM integration + why For each: free tier available? Monthly cost at 1000 users? Alternative if this fails? Final output: starter repo structure + estimated monthly cost at 3 scales (100 / 1K / 10K users).
β–Ά Show alternate versions
Alt A β€” Zero-code / low-code stack
Recommend a no-code / low-code stack for [PRODUCT] for a non-technical founder with $200/month budget. Tools: Bubble / Webflow / Supabase / Zapier / Make / Airtable / etc. Speed over scalability. What can be live in 2 weeks?
Alt B β€” Minimum infra stack (under $50/month)
Design a tech stack for [PRODUCT] that costs under $50/month to run, handles up to 500 active users, and can be maintained by 1 developer. Prioritize: managed services, free tiers, serverless, and simplicity over performance.
PROMPT T-02
System Architecture Design
Design the full system before writing any code.
Design the system architecture for: Product: [PRODUCT] Stack: [T-01 OUTPUT] Core features: [TOP FEATURES] Output: 1. Component diagram (describe all services and how they connect) 2. Data flow: trace a user's full journey from login to core action 3. API structure: list all endpoints (method, path, purpose, auth required) 4. Database tables: name, key fields, relationships 5. Third-party integrations and their role 6. Security boundaries (what is public vs private) 7. Caching strategy 8. Failure points and fallback plan for each Format the API list as a table. Format DB tables as schema definitions.
PROMPT T-03
Database Schema Design
Design a complete, normalized database from scratch.
Design a complete database schema for: Product: [PRODUCT] Database: [POSTGRES / MYSQL / MONGODB] Core entities: [LIST MAIN DATA OBJECTS] For each table: - Table name - All columns with types, constraints, defaults - Primary key, foreign keys, indexes - Relationships (1:1, 1:many, many:many) Rules: - Normalize to 3NF minimum - Include: created_at, updated_at, soft delete (deleted_at) on all main tables - Multi-tenancy support (if SaaS): org_id on all tenant-scoped tables - Include Row Level Security policies if using Supabase Output as SQL CREATE TABLE statements + an ERD description. Flag any design decisions that involve tradeoffs.
PROMPT T-04
AI/LLM Integration Architecture
Design how Claude/GPT plugs into the product correctly.
AI-SPECIFIC
Design the AI integration layer for: Product: [PRODUCT] AI model to use: [Claude Sonnet / GPT-4o / Gemini] AI features needed: [LIST AI FEATURES] Design: 1. Prompt architecture (system prompt structure, context injection, user input handling) 2. Context window management (what goes in, what gets summarized, what is cut) 3. Streaming vs batch (which features need streaming output?) 4. Cost control (estimate tokens per feature, monthly cost at 1K users) 5. Caching strategy (what AI outputs can be cached?) 6. Fallback plan (if API is down or rate limited) 7. Evaluation loop (how do we know if AI responses are good?) 8. Safety filters (what outputs do we block or review?) Output the system prompt template for the core AI feature. Include [VARIABLES] in caps.
PROMPT T-05
Security Review Checklist
Security audit before launch β€” non-negotiable.
REQUIRED
Perform a security review for: Product: [PRODUCT] Stack: [STACK] Data stored: [USER DATA TYPES] Review and give pass/fail + fix for each: Auth & Access: β–‘ JWT expiry + refresh strategy β–‘ Rate limiting on auth endpoints β–‘ Password hashing (bcrypt/argon2) β–‘ OAuth token storage β–‘ RBAC / permissions model Data Security: β–‘ SQL injection prevention β–‘ XSS prevention β–‘ CSRF protection β–‘ Sensitive data encryption at rest β–‘ Secure API key management β–‘ PII data handling compliance Infrastructure: β–‘ HTTPS everywhere β–‘ Env vars vs hardcoded secrets β–‘ Dependency vulnerability scan β–‘ Logging (no PII in logs) β–‘ Backup strategy Output: Priority-ordered fix list with code snippets for critical issues.
Layer 4 β€” Business Model + GTM
Define how the product makes money and how the first 100 customers find it.
PROMPT B-01
Revenue Model Design
Design the complete monetization architecture.
CRITICAL
You are a monetization strategist. Design the revenue model for: Product: [PRODUCT] Personas: [ICP 1, ICP 2, ICP 3] Competitors' pricing: [FROM V-02] Our differentiation: [UNIQUE ANGLE] Design: 1. Primary revenue model (subscription / usage / marketplace / freemium / one-time) 2. Pricing tiers (3 recommended): name, price, what's included, who it's for 3. Free tier scope (what's free and why β€” acquisition or goodwill?) 4. Upsell triggers (what feature makes users upgrade?) 5. Annual vs monthly discount strategy 6. Enterprise/custom tier criteria 7. Expansion revenue opportunities (seats, usage, add-ons) Calculate unit economics: - ARPU target (average revenue per user) - Target gross margin % - Payback period assumption - MRR needed to cover $[MONTHLY COSTS] in operating costs Output a pricing page copy draft for each tier.
β–Ά Show alternate versions
Alt A β€” Usage-based pricing model
Design a usage-based pricing model for [PRODUCT]. What is the billing unit (API calls / seats / documents / outputs)? Set pricing per unit, include a free quota, and design overage pricing. Calculate: revenue at 100 / 500 / 1000 active users.
Alt B β€” Marketplace / commission model
Design a marketplace revenue model for [PRODUCT]. What % commission is defensible? What services can we add (featured listings, subscriptions)? How do we bootstrap supply and demand simultaneously? When does the commission model break and what do we switch to?
PROMPT B-02
Go-To-Market Strategy
The exact plan to get first 100 paying customers.
You are a GTM expert. Create a go-to-market plan for: Product: [PRODUCT] Primary ICP: [PERSONA 1] Budget: [MARKETING BUDGET] Goal: First 100 paying customers in [TIMEFRAME] Design: 1. Launch channels (rank by: low cost, fast feedback, right audience) 2. Week 1–2: Pre-launch (waitlist, community seeding, beta recruitment) 3. Week 3–4: Launch (Product Hunt, HN, Reddit, newsletters, cold outreach) 4. Month 2–3: Growth (content, SEO foundations, partnership, paid experiments) For each channel: - Platform + specific community/subreddit/newsletter - Message/angle that will resonate - Expected CAC - Expected volume Cold outreach template (50-word max email that gets replies): Landing page headline options (3 variants A/B test): First 10 beta users acquisition plan:
PROMPT B-03
Financial Model (Lean)
12-month P&L projection with assumptions.
Build a 12-month financial model for: Product: [PRODUCT] Pricing: [TIERS + PRICES FROM B-01] Monthly costs: [INFRA + TEAM + TOOLS] Growth assumption: [% MoM USER GROWTH] Churn assumption: [MONTHLY CHURN %] Free-to-paid conversion: [%] Output a month-by-month table: | Month | New Users | Paid Users | MRR | Churn | Net MRR | Costs | Profit/Loss | Key milestones to highlight: - Month we hit default alive (revenue > costs) - Month we hit $1K MRR / $10K MRR - Total capital needed before profitability Assumptions section: list every assumption made β€” flag the 3 most dangerous ones.
PROMPT B-04
Positioning & Messaging
Nail the one-liner, tagline, and core messaging pillars.
You are a brand strategist. Create the messaging framework for: Product: [PRODUCT] Category: [WHAT WE DO] Primary user: [ICP] Key benefit: [TOP VALUE PROP] Competitors: [MAIN ALTERNATIVES] Create: 1. Positioning statement: "For [user] who [problem], [product] is the [category] that [key benefit]. Unlike [competitor], we [key differentiator]." 2. Tagline options (5 variants, 3–7 words each) 3. One-liner (25 words max) for cold outreach 4. Hero headline for landing page (pain-led vs outcome-led variants) 5. Three messaging pillars (core claims we can prove) 6. Proof points for each pillar (stats, case studies, testimonials) 7. Words to avoid (category clichΓ©s that make us sound like everyone else) 8. Tone of voice: 3 adjectives, 1 brand we sound like
Layer 5 β€” Build Engine
Prompts for generating production-grade code, components, and deployment configs.
PROMPT BL-01
Full-Stack Feature Build
Build any feature end-to-end: DB β†’ API β†’ UI in one prompt.
POWER PROMPT
You are a senior full-stack engineer. Build this feature completely: Feature: [FEATURE NAME] Stack: [STACK FROM T-01] User story: [FROM P-04] DB schema: [RELEVANT TABLES FROM T-03] Build in this order: 1. DB migration (SQL) 2. API route(s) β€” full code with error handling 3. Input validation schema (Zod / Joi) 4. Frontend component β€” full code, responsive, accessible 5. Loading + error + empty states for the UI 6. Unit test for the core logic 7. Integration test for the API route Rules: - TypeScript only - Handle all error cases explicitly (no silent failures) - Include JSDoc comments on public functions - No TODO comments β€” write the actual implementation - Production-ready: auth check, rate limit, input sanitize Output each file as a separate code block with the file path as a comment.
β–Ά Show alternate versions
Alt A β€” Frontend only (UI component)
Build a React component for [FEATURE]. Use [Tailwind / shadcn / MUI]. Include: props interface, loading state, error state, empty state, responsive breakpoints. Add micro-animations using Framer Motion for key interactions. Follow WCAG 2.1 AA accessibility standards.
Alt B β€” Backend API only
Build a REST API for [FEATURE] using [Express/FastAPI/Next.js API routes]. Include: auth middleware, request validation, error handling, response formatting, rate limiting. Write OpenAPI spec for each endpoint. Include curl examples for testing.
PROMPT BL-02
Auth System Build
Complete auth: signup, login, sessions, OAuth, RBAC.
Build a complete authentication system for: Stack: [FRAMEWORK] Auth provider: [Supabase Auth / Clerk / NextAuth / Custom JWT] OAuth providers needed: [Google / GitHub / etc] User roles needed: [admin / user / etc] Build: 1. Sign up flow (email + OAuth) with email verification 2. Sign in flow with remember me 3. Password reset flow 4. Session management + JWT refresh 5. Auth middleware (protect routes) 6. Role-based access control (RBAC) setup 7. "Magic link" login option Include: - Database tables for users, sessions, roles - All API routes with validation - Frontend forms with client-side validation - Error handling for all auth failure cases - Rate limiting on auth endpoints
PROMPT BL-03
Payments Integration
Stripe setup: one-time, subscriptions, webhooks, portal.
Integrate Stripe payments for: Product: [PRODUCT] Payment type: [SUBSCRIPTION / ONE-TIME / USAGE-BASED] Pricing tiers: [FROM B-01] Stack: [FRAMEWORK] Build: 1. Stripe product + price setup (CLI commands) 2. Checkout session creation API route 3. Success + cancel redirect handling 4. Webhook handler for: payment_intent.succeeded, customer.subscription.*, invoice.payment_failed 5. Customer portal integration (manage subscription) 6. Subscription status check middleware 7. DB schema: subscriptions, payments tables 8. Frontend: pricing page, upgrade flow, billing section Include: test mode setup, webhook local testing with Stripe CLI, error handling for failed payments.
PROMPT BL-04
Landing Page Generator
Build a high-converting landing page from your positioning doc.
Build a high-converting landing page for: Product: [PRODUCT] Positioning: [FROM B-04 β€” headline, tagline, pillars] Primary CTA: [SIGN UP FREE / REQUEST DEMO / JOIN WAITLIST] Social proof: [TESTIMONIALS / USER COUNT / LOGOS] Build in Next.js + Tailwind with these sections: 1. Hero: headline, subheadline, CTA, hero image/mockup area 2. Social proof bar (logos or user count) 3. Problem section (pain before our solution) 4. Solution section (3 key features with icons) 5. How it works (3-step visual flow) 6. Testimonials (3 cards) 7. Pricing section (3 tiers) 8. FAQ (5 questions) 9. Final CTA + footer SEO: Add meta tags, OG image placeholder, semantic HTML. Performance: lazy load images, no layout shift, Lighthouse 90+.
PROMPT BL-05
Deployment & DevOps Setup
CI/CD, environment management, monitoring from day 1.
Set up the DevOps infrastructure for: Stack: [STACK] Hosting: [Vercel / Railway / Render / AWS / GCP] DB: [DATABASE + HOSTING] Set up: 1. Environment management (.env.local, .env.production, secrets) 2. GitHub Actions CI pipeline: lint β†’ test β†’ build β†’ deploy 3. Preview deployments for PRs 4. Production deployment with zero-downtime 5. Database migration strategy (auto-run in CI vs manual) 6. Error monitoring setup (Sentry β€” free tier config) 7. Uptime monitoring (Better Uptime / UptimeRobot β€” free) 8. Log aggregation setup 9. Backup strategy for database 10. Custom domain + SSL setup Output: all config files, GitHub Actions YAML, and Dockerfile if needed.
Client Loop β€” Keeping Them In The Building
Prompts for discovery, updates, feedback, and managing expectations at every stage.
CLIENT-01
Discovery Session Framework
Run the first client session and extract everything you need to start.
SESSION 1
You are a product discovery expert conducting the first session with a new client. Their raw idea: [CLIENT'S WORDS β€” PASTE EXACTLY AS THEY SAID IT] Ask these questions in a warm, conversational way β€” not as a list. Explore one at a time, go deep, don't rush: 1. Tell me about the moment you first thought of this. What were you frustrated about? 2. Who do you imagine using this first? Do you know them personally? 3. Have you talked to potential users yet? What did they say? 4. What have you already tried building or testing? 5. What does success look like in 12 months? (revenue, users, impact) 6. What would make you declare this a failure and shut it down? 7. Who else is working on this with you? 8. What's your biggest fear about this project? 9. What resources do you have? (time, money, team, connections) 10. What do you need from me specifically? After all answers: output the Client Brief document to be passed into Prompt V-01.
CLIENT-02
Weekly Status Update Generator
Generate a clear, professional progress update for clients.
Generate a weekly status update email for my client. Week number: [N] What we planned: [PASTE LAST WEEK'S PLAN] What we actually did: [BULLET LIST OF COMPLETED ITEMS] Blockers encountered: [LIST] Next week plan: [BULLET LIST] Key decision needed from client: [IF ANY] Current status: [ON TRACK / AT RISK / DELAYED] Write a professional but warm email that: - Opens with 1 positive highlight - Is honest about blockers without causing panic - Clearly states what needs client input - Closes with a clear ask for the week - Is max 200 words Also generate: a 3-slide deck outline (title, progress, next week) for async Loom update.
CLIENT-03
Milestone Review & Demo Script
Prepare for milestone demos with talking points and expected questions.
Prepare a milestone demo script for: Milestone: [MILESTONE NAME, e.g. "MVP Demo - Week 4"] What was built: [FEATURE LIST] What still has gaps: [HONEST LIST] Client expectations set: [WHAT WE PROMISED] Next milestone: [WHAT COMES NEXT] Write: 1. Demo flow (step-by-step what to show and in what order) 2. Talking points for each feature (benefit-first, not tech-first) 3. How to handle the gaps gracefully (frame as "by design" or "in backlog") 4. 5 questions the client will likely ask β€” and ideal answers for each 5. Call-to-action at end of demo (what decision do we need from them?) 6. Follow-up email template (send within 1 hour of demo)
CLIENT-04
Scope Change Management
Handle "can we add just one more thing" professionally.
A client wants to add a new feature mid-project. Help me respond. Original scope: [ORIGINAL FEATURES] New request: [WHAT CLIENT WANTS TO ADD] Current timeline: [WEEKS REMAINING] Impact: [ROUGH ESTIMATE OF ADDED WORK] Write: 1. A professional email acknowledging the request positively 2. Analysis: what is the actual impact on timeline + cost? 3. Three options to offer the client: a. Add it β€” here's the cost + timeline impact b. Add it in Phase 2 β€” here's when c. Replace something β€” what to cut instead 4. Recommended option with rationale 5. A scope change document template they can sign off on Keep the tone: collaborative, not defensive. We're on their side.
CLIENT-05
User Feedback Analysis
Turn messy user feedback into a prioritized action plan.
Analyze this batch of user feedback and turn it into action: Product: [PRODUCT] Feedback collected: [PASTE ALL RAW FEEDBACK β€” INTERVIEWS, EMAILS, REVIEWS] Current version: [WHAT'S LIVE] Analyze: 1. Cluster feedback into themes (group similar issues) 2. Frequency + sentiment for each theme 3. Distinguish: bug reports vs feature requests vs UX confusion vs missing features 4. Quote the most impactful individual piece of feedback (the one that should keep us up at night) 5. Prioritized action list: Critical / High / Medium / Low 6. Feature requests to add to backlog with user quote as rationale 7. Quick wins (can be fixed in < 1 day) β€” ship these first 8. What the data tells us about our ICP fit (are the right people using this?)
Complete Prompt Library
Every prompt in the system, organized. Use as a quick-reference during any session.
πŸ’‘ Validation Prompts (V-01 to V-05)β–Ά
  • β–Έ
    V-01 β€” Idea extraction & structuring
  • β–Έ
    V-02 β€” Market size & TAM/SAM/SOM
  • β–Έ
    V-03 β€” Competitor deep-dive & positioning map
  • β–Έ
    V-04 β€” User persona generation (3 ICPs)
  • β–Έ
    V-05 β€” Feasibility scorecard (GO/NO-GO)
πŸ“‹ Product Planning Prompts (P-01 to P-05)β–Ά
  • β–Έ
    P-01 β€” Full PRD generation
  • β–Έ
    P-02 β€” MVP scope definition
  • β–Έ
    P-03 β€” 3-phase roadmap generation
  • β–Έ
    P-04 β€” User story generator
  • β–Έ
    P-05 β€” Feature prioritization (RICE)
βš™οΈ Tech Architecture Prompts (T-01 to T-05)β–Ά
  • β–Έ
    T-01 β€” Tech stack recommendation
  • β–Έ
    T-02 β€” System architecture design
  • β–Έ
    T-03 β€” Database schema design
  • β–Έ
    T-04 β€” AI/LLM integration architecture
  • β–Έ
    T-05 β€” Security review checklist
πŸ’° Business Model Prompts (B-01 to B-04)β–Ά
  • β–Έ
    B-01 β€” Revenue model design
  • β–Έ
    B-02 β€” Go-to-market strategy
  • β–Έ
    B-03 β€” Financial model (12-month P&L)
  • β–Έ
    B-04 β€” Positioning & messaging
πŸ”¨ Build Engine Prompts (BL-01 to BL-05)β–Ά
  • β–Έ
    BL-01 β€” Full-stack feature build
  • β–Έ
    BL-02 β€” Auth system build
  • β–Έ
    BL-03 β€” Stripe payments integration
  • β–Έ
    BL-04 β€” Landing page generator
  • β–Έ
    BL-05 β€” Deployment & DevOps setup
🀝 Client Loop Prompts (C-01 to C-05)β–Ά
  • β–Έ
    C-01 β€” Discovery session framework
  • β–Έ
    C-02 β€” Weekly status update generator
  • β–Έ
    C-03 β€” Milestone demo script
  • β–Έ
    C-04 β€” Scope change management
  • β–Έ
    C-05 β€” User feedback analysis
Master Context Prompt
Paste this at the start of EVERY new Claude session to maintain continuity.
You are my AI co-founder helping me build [PRODUCT NAME]. Here is our complete product context: IDEA BRIEF: [PASTE V-01 OUTPUT] PRIMARY PERSONA: [PASTE V-04 PERSONA 1] CURRENT PHASE: [VALIDATION / PLANNING / BUILD / LAUNCH] STACK: [PASTE T-01 OUTPUT] WHAT WE'VE BUILT: [BULLET LIST] WHAT'S WORKING: [USER FEEDBACK HIGHLIGHTS] TOP PRIORITY THIS WEEK: [THIS WEEK'S GOAL] Today I need help with: [SPECIFIC TASK] Do not repeat back this context. Jump straight into helping.

Bonus Prompts
Email Campaign Generator
Write a 5-email onboarding sequence for [PRODUCT]. Day 0: Welcome + get started Day 2: Show the #1 feature Day 5: Social proof + use case Day 10: Upgrade prompt Day 14: Check-in / survey Each email: subject line (A/B variant), 150-word max body, 1 clear CTA.
Product Hunt Launch Copy
Write Product Hunt launch copy for [PRODUCT]. Include: tagline (60 chars max), description (260 chars), long description (500 words), first comment from maker (personal story + ask for support), and 5 questions to ask the community for engagement.
Technical Interview (Hire Engineers)
Create a take-home technical challenge to hire a [ROLE] for [PRODUCT]. Relevant stack: [STACK]. Time limit: 3 hours. Include: problem statement, requirements, bonus points, evaluation rubric.
Investor Pitch Deck Outline
Create a 10-slide pitch deck outline for [PRODUCT]. Context: [PASTE ALL OUTPUTS SO FAR] Slides: Problem, Solution, Market, Product demo, Business model, Traction, Team, Roadmap, Ask, Contact. For each slide: title, 3 bullet points, 1 visual suggestion, 1 key number to feature.
Tools Arsenal
Every tool you need to run the full ecosystem. Free-first. Paid when it matters.
πŸ€– AI & LLM
ToolPurposeCostBest For
Claude Sonnet 4.5Main orchestration, code, strategyEverything in this handbook
Claude Opus 4.5Complex reasoning, architectureT-02, T-03, B-03 prompts
GPT-4oSecond opinion, multimodalCross-validate key decisions
Perplexity ProMarket research with web searchV-02, V-03 prompts
Whisper (OpenAI)Transcribe client callsFree (local)Capture discovery sessions
πŸ— Infra & Hosting
ToolCost
Vercel
Frontend hosting + preview deploys
Free tier
Railway
Backend + DB hosting
Supabase
Postgres + auth + storage + realtime
Free tier
Cloudflare
DNS, CDN, Workers edge
Free tier
Upstash
Redis cache + rate limiting
Free tier
πŸ’³ Payments & Revenue
ToolCost
Stripe
Payments, subscriptions, billing
Lemon Squeezy
Merchant of record (no tax headache)
Paddle
B2B SaaS billing + tax
Stripe Billing
Usage-based billing
πŸ“§ Email & Comms
ToolCost
Resend
Transactional email API
3K/month free
Loops
SaaS email marketing
Free to 1K
Postmark
High deliverability transactional
100/month free
Crisp
Live chat + support
Free tier
πŸ“Š Analytics & Monitoring
ToolCost
PostHog
Product analytics + session replay
1M events free
Sentry
Error tracking + performance
Free tier
Plausible
Privacy-first web analytics
Better Uptime
Uptime monitoring + status page
Free tier
πŸ›  Dev Tools
ToolCost
GitHub
Version control + CI/CD + Actions
Free
Linear
Issue tracking + sprint planning
Free to 3 users
Cursor
AI-native code editor
Prettier + ESLint
Code formatting + linting
Free (OSS)
πŸ“ Docs & Collaboration
ToolCost
Notion
Product docs, PRDs, wikis
Free for personal
Loom
Async video updates for clients
Free tier
Figma
Design, prototyping, handoff
Free tier
Excalidraw
Quick diagrams + whiteboarding
Free (OSS)
Minimum Viable Stack (MVC β€” Minimum Viable Cost)
Next.js (free) Supabase (free) Vercel (free) Resend (free) PostHog (free) Stripe (% only) Claude API (~$50/mo) Domain ($12/yr)
Total fixed cost to launch: ~$62/month until you hit revenue.