---
title: "International Product Budget Plan"
version: "1.1"
last_updated: "2026-07-18"
status: "public"
language: "en"
---

# Budget Plan — International AI Creative-Tools Platform

> **Document type:** Internal financial and go-to-market plan  
> **Prepared:** 1 July 2026  
> **Updated:** 18 July 2026  
> **Scope:** A separately branded, English-language AI creative-tools SaaS for global customers, initially focused on the US and Europe. It is not the Iranian Luma product and does not define Luma’s local pricing or legal obligations.  
> **Currency:** USD unless stated otherwise.

The figures in this document are planning assumptions, not vendor quotations. Hosting, model inference, payment, advertising and contractor rates change frequently and must be revalidated before procurement or launch.

---

## Table of contents

1. [Executive summary](#1-executive-summary)
2. [Decisions required before spending](#2-decisions-required-before-spending)
3. [Product and launch scope](#3-product-and-launch-scope)
4. [One-time build and launch costs](#4-one-time-build-and-launch-costs)
5. [Fixed monthly operating costs](#5-fixed-monthly-operating-costs)
6. [Variable cost of goods sold](#6-variable-cost-of-goods-sold)
7. [International pricing and plans](#7-international-pricing-and-plans)
8. [Free access and trial strategy](#8-free-access-and-trial-strategy)
9. [Marketing and customer acquisition](#9-marketing-and-customer-acquisition)
10. [Creator, affiliate and partnership growth](#10-creator-affiliate-and-partnership-growth)
11. [Budget scenarios](#11-budget-scenarios)
12. [Unit economics and break-even](#12-unit-economics-and-break-even)
13. [Ninety-day launch plan](#13-ninety-day-launch-plan)
14. [Risk register](#14-risk-register)
15. [Financial controls and reporting](#15-financial-controls-and-reporting)
16. [Open decisions](#16-open-decisions)

---

## 1. Executive summary

The business has four major cost groups:

| Cost group | Planning assumption |
| :--- | :--- |
| One-time product build | Approximately $5k–15k for a founder-led build, $30k–60k with freelancers, or $30k–75k with an agency |
| Fixed monthly overhead | Approximately $65–165 per month at a very lean start and $560–1,300 per month at growth stage, excluding payroll |
| Variable COGS | AI inference, media storage and delivery, payment processing, refunds and failed-generation leakage |
| Growth spend | Paid media, creators, affiliate commissions, content and partnerships; normally the largest discretionary cash requirement |

### Six-month capital scenarios

| Scenario | Up-front build and setup | Monthly burn | Six-month planning range |
| :--- | :--- | :--- | :--- |
| Lean bootstrap | $15k–35k | $8k–12k | $60k–100k |
| Moderate growth | $45k–80k | $20k–30k | $170k–260k |
| Aggressive scale | $80k–150k | $50k–75k | $380k–600k |

The recommended starting point is a lean validation phase with strict usage metering, limited paid acquisition and a product scope narrow enough to establish retention before scale.

---

## 2. Decisions required before spending

The following decisions have more impact than most individual cost items:

### 2.1. Build path

- founder-led implementation;
- small freelance team;
- dedicated product team;
- agency build;
- hybrid reuse of Luma knowledge with a new codebase and brand.

### 2.2. Monetization model

- hard paywall;
- paid subscription with credit allowance;
- subscription plus pay-as-you-go top-ups;
- capped freemium;
- credit-card trial;
- reverse trial that temporarily unlocks premium features.

### 2.3. Free usage

Open free credits can create abuse, high COGS and low-intent users. The recommended experiment is a tightly capped trial with lower-cost models, output restrictions and anti-abuse controls.

### 2.4. Merchant of Record

A Merchant of Record can simplify tax collection, invoicing and international compliance, but normally charges more than direct payment processing. The decision should be based on target jurisdictions, team capacity and expected volume.

### 2.5. Target segment

The initial product should not target every AI user. One or two segments must be selected for launch, such as:

- ecommerce product-content teams;
- creators and design agencies;
- fashion and apparel merchants;
- video-first social media creators;
- developers needing several AI models behind one account.

---

## 3. Product and launch scope

### 3.1. Initial product set

The proposed product mirrors the strongest parts of Luma’s Creativity Studio:

- AI image generation;
- AI image editing;
- background removal;
- image upscaling;
- virtual clothing try-on;
- text-to-video;
- image-to-video;
- reference-to-video;
- multi-model AI chat;
- credit wallet and usage history;
- subscription and top-up billing;
- asset library;
- support and basic account administration.

Workflow, Smart Assistant, team collaboration, public galleries and marketplace capabilities should be treated as later phases unless a validated launch segment requires them.

### 3.2. Minimum launch requirements

- authentication and account recovery;
- secure credit ledger;
- payment and subscription handling;
- model routing and provider failover rules;
- usage metering;
- generation status and retry handling;
- asset storage and deletion;
- rate limiting and abuse prevention;
- support process;
- privacy, terms and acceptable-use documents;
- product analytics and error monitoring;
- financial reconciliation.

### 3.3. Explicit exclusions for the first release

- broad no-code automation platform;
- full enterprise identity suite;
- unrestricted public marketplace;
- complex team permissions;
- native mobile applications;
- custom model training;
- guaranteed SLA;
- unlimited usage plans.

---

## 4. One-time build and launch costs

### 4.1. Product development

Assumed complexity: a credit-based SaaS using third-party image, video and language-model APIs, with billing, asset storage and a production user interface.

| Delivery path | Low | Typical | High | Notes |
| :--- | ---: | ---: | ---: | :--- |
| Founder-led / starter kit | $1k | $5k–15k | $30k | Lowest cash cost; highest founder-time requirement |
| Freelance developers | $16k | $30k–60k | $120k+ | Depends on geography, seniority and project management |
| Agency | $15k–30k | $30k–75k | $150k+ | Faster coordination, but high cost and lock-in risk |
| Small internal team | $25k | $60k–120k | $200k+ | Better long-term ownership; includes recruitment and payroll runway |

Recommended bootstrap range: $15k–40k, assuming substantial reuse of product knowledge, UI patterns and provider integrations.

### 4.2. Brand and design

| Item | Lean range | Professional range |
| :--- | ---: | ---: |
| Naming and domain research | $0–500 | $1k–3k |
| Visual identity | $0–1k | $3k–10k |
| Product design and prototypes | $1k–5k | $10k–30k |
| Marketing site | $500–3k | $5k–20k |
| Initial creative library | $500–2k | $5k–15k |

In-house design capability can reduce cash cost, but brand, product and conversion design still require dedicated time.

### 4.3. Legal and company setup

Planning allowance:

- entity formation: $500–3k;
- accounting and banking setup: $500–2k;
- Terms, Privacy, Acceptable Use and Data Processing documents: $1k–8k;
- trademark search and filing: $500–3k per priority jurisdiction;
- contractor and affiliate agreements: $500–2k.

Legal cost depends heavily on jurisdiction and business structure.

### 4.4. Domain and launch assets

- `.com`: approximately $10–20 per year;
- `.ai`: often materially more expensive and may require multi-year registration;
- email and basic brand accounts: $100–500 initial setup;
- launch video, demos and screenshots: $1k–10k depending on production quality.

---

## 5. Fixed monthly operating costs

### 5.1. Infrastructure

Recommended principle: generated media should use storage and delivery with predictable egress economics. Large video traffic can make conventional object-storage egress expensive.

| Stage | Monthly infrastructure assumption |
| :--- | ---: |
| Prototype | $5–50 |
| Early launch | $50–250 |
| Growth | $250–1,500 |
| High media volume | Usage-based; can exceed $5k without storage and delivery controls |

Infrastructure includes:

- frontend hosting;
- backend and queue workers;
- database;
- object storage;
- media delivery;
- logs and monitoring;
- backups;
- analytics events;
- email and notifications.

### 5.2. Supporting SaaS tools

| Category | Lean monthly assumption | Growth assumption |
| :--- | ---: | ---: |
| Authentication | $0–25 | $25–200 |
| Transactional email | $0–20 | $50–300 |
| Product analytics | $0–50 | $100–500 |
| Error monitoring | $0–30 | $50–300 |
| Customer support | $0–100 | $200–1,500 |
| Affiliate platform | $49–100 | $100–500 |
| Accounting / finance tools | $20–100 | $100–500 |
| Design and collaboration | $20–100 | $100–500 |

### 5.3. Payroll and contractors

Payroll is excluded from the small fixed-overhead figures above and must be budgeted separately.

Example lean monthly team:

| Role | Monthly contractor / distributed-team range |
| :--- | ---: |
| Full-stack / platform engineer | $4k–10k |
| AI integration engineer | $4k–12k |
| Product designer | $2k–7k |
| Growth/content operator | $2k–8k |
| Support / operations | $1k–4k |

A founder-led launch can reduce cash payroll but increases execution and key-person risk.

---

## 6. Variable cost of goods sold

### 6.1. AI inference

Every generation creates provider cost. Representative planning ranges:

#### Images

| Type | Raw-cost planning range |
| :--- | ---: |
| Budget image generation | $0.003–0.02 |
| Standard image generation or editing | $0.02–0.08 |
| Premium image generation | $0.10–0.30+ |
| Background removal | $0.001–0.01 |
| Upscaling | $0.004–0.05 |
| Virtual try-on | $0.04–0.15 |

#### Video

| Type | Raw-cost planning range |
| :--- | ---: |
| Budget short clip | $0.50–1.50 |
| Standard short clip | $1.50–4.00 |
| Premium clip | $3.00–8.00+ |

Video is the largest COGS and abuse risk. It requires strict credit multipliers, duration caps and retry controls.

#### Text and chat

Language-model cost depends on input tokens, output tokens, context length, caching and model class. Long conversations can increase input cost on every turn.

### 6.2. Markup policy

A target retail price of approximately three to four times raw inference cost is a reasonable starting assumption, but the true required multiplier must include:

- failed-generation leakage;
- retries and support credits;
- storage and delivery;
- payment fees;
- affiliate commission;
- refunds and fraud;
- taxes handled by the business;
- customer support;
- model-price volatility.

### 6.3. Payment processing

Plan for an effective 5–8% revenue cost for international payments after transaction fees, fixed charges, foreign exchange, refunds and Merchant-of-Record costs.

### 6.4. Refunds, disputes and fraud

Budget assumptions:

- refunds: 1–3% of gross revenue;
- payment disputes: 0.3–1%;
- promotional and support credits: 1–3%;
- free-tier abuse: potentially much higher without controls.

### 6.5. Target gross margin

A realistic blended gross-margin target for a multi-model AI product is approximately 50–65%, rather than traditional software margins, especially when video is a major use case.

---

## 7. International pricing and plans

### 7.1. Pricing principles

- meter expensive models accurately;
- separate subscription allowance from paid top-ups;
- avoid unlimited generation claims;
- show credit cost before execution;
- make video economics explicit;
- preserve a path for occasional users;
- use annual plans only after monthly retention is proven;
- keep plan differences understandable.

### 7.2. Initial plan hypothesis

| Feature | Starter | Creator | Pro | Studio |
| :--- | :---: | :---: | :---: | :---: |
| Monthly price | $12 | $29 | $69 | $149 |
| Included credits | 1,000 | 3,000 | 8,000 | 20,000 |
| Top-up discount | 0% | 5% | 10% | 15% |
| Concurrent jobs | 1 | 3 | 6 | 15 |
| Storage | 5 GB | 25 GB | 100 GB | 500 GB |
| Brand presets | 3 | 15 | 50 | Unlimited |
| Support | Standard | Standard | Priority | Priority |

These figures are hypotheses and must be recalculated using actual provider cost and observed user mix.

### 7.3. Pay-as-you-go

Recommended options:

- minimum top-up: $10;
- larger packs with limited volume discount;
- no hidden expiration for purchased credits unless clearly disclosed;
- separate rules for promotional credits;
- clear credit ledger and transaction history.

### 7.4. Annual pricing

Do not launch large annual discounts before churn and support costs are known. A 15–20% annual discount can be tested after monthly cohorts show stable retention.

---

## 8. Free access and trial strategy

### 8.1. Risk of open free credits

- bot and multi-account abuse;
- immediate video COGS;
- low-intent signups;
- payment conversion distortion;
- support load;
- poor data quality;
- output reselling by non-paying users.

### 8.2. Recommended trial experiment

- short reverse trial or capped allowance;
- lower-cost image and chat models;
- no premium video or strict video cap;
- watermark or reduced resolution where appropriate;
- account and device abuse checks;
- no cash value or transfer;
- clear expiry;
- conversion prompt after demonstrated value.

### 8.3. Experiments

Test separately:

- no-card trial;
- card-required trial;
- one-time starter credits;
- daily capped allowance;
- demo gallery without generation;
- first-generation free after email verification;
- free images but paid video.

Primary metric: paid conversion adjusted for COGS and thirty-day retention, not raw signup volume.

---

## 9. Marketing and customer acquisition

### 9.1. Paid acquisition planning

| Stage | Monthly paid-media assumption |
| :--- | ---: |
| Validation | $2k–5k |
| Early scale | $5k–15k |
| Growth | $20k–50k |
| Aggressive expansion | $50k+ |

Paid acquisition should not scale before activation, checkout conversion and retention are measurable.

### 9.2. Channel order

1. Search demand for high-intent use cases;
2. retargeting;
3. creator content and whitelisting;
4. affiliate traffic;
5. use-case landing pages and SEO;
6. social prospecting;
7. partnerships and integrations.

### 9.3. Content budget

Lean monthly content allowance:

- editorial and SEO: $1k–4k;
- short-form video and design: $1k–5k;
- demos and case studies: $500–3k;
- localization: $500–2k per target language.

### 9.4. CAC expectations

CAC cannot be estimated from clicks alone. It must include:

- media spend;
- creator fees;
- affiliate commission;
- trial COGS;
- sales and support time;
- refunds and disputes;
- non-converting free usage.

Initial planning ranges should be treated as hypotheses:

- self-serve paid user: $40–150;
- premium creator or small team: $100–400;
- sales-assisted business account: $500–5,000+.

---

## 10. Creator, affiliate and partnership growth

### 10.1. Creator programme

Recommended structure:

- small fixed production fee;
- tracked link or code;
- commission on net paid revenue;
- bonus for retained customers;
- usage rights for successful content;
- product-claim review before publication.

### 10.2. Affiliate economics

Possible starting structure:

- 15–25% of first-year net subscription revenue;
- or 20–30% recurring for a limited period;
- lower commission on credit top-ups with high COGS;
- hold period for refunds and disputes;
- anti-self-referral rules;
- payout threshold.

The final rate must be modelled against gross margin.

### 10.3. Partnership targets

- design communities;
- ecommerce platforms;
- marketing agencies;
- creator-education businesses;
- template marketplaces;
- stock-asset providers;
- hosting and developer platforms;
- fashion and product-photography communities.

---

## 11. Budget scenarios

### 11.1. Lean bootstrap

| Item | Up-front | Monthly |
| :--- | ---: | ---: |
| Product build | $10k–25k | — |
| Brand, legal and setup | $5k–10k | — |
| Infrastructure and SaaS | — | $200–700 |
| Contractors | — | $3k–6k |
| Paid media | — | $2k–5k |
| Creators and content | — | $1k–3k |
| COGS reserve | — | $1k–3k |
| Contingency | $3k–5k | $500–1k |

Approximate monthly burn: $8k–12k.

### 11.2. Moderate growth

| Item | Up-front | Monthly |
| :--- | ---: | ---: |
| Product build | $30k–60k | — |
| Brand, legal and setup | $10k–20k | — |
| Infrastructure and SaaS | — | $1k–3k |
| Team and contractors | — | $8k–15k |
| Paid media | — | $8k–15k |
| Creators and content | — | $3k–7k |
| COGS reserve | — | $3k–8k |
| Contingency | $5k–10k | $1k–2k |

Approximate monthly burn: $20k–30k.

### 11.3. Aggressive scale

| Item | Up-front | Monthly |
| :--- | ---: | ---: |
| Product build and team setup | $60k–120k | — |
| Brand, legal and localisation | $20k–40k | — |
| Infrastructure and SaaS | — | $3k–10k |
| Team | — | $20k–35k |
| Paid media | — | $20k–50k |
| Creators and partnerships | — | $8k–20k |
| COGS reserve | — | $8k–20k |
| Contingency | $10k–20k | $3k–5k |

Approximate monthly burn: $50k–75k.

---

## 12. Unit economics and break-even

### 12.1. Core formulas

```text
Net Revenue = Gross Revenue - Taxes collected - Refunds - Payment fees

Gross Profit = Net Revenue - AI COGS - Storage and delivery - Support credits

Contribution Margin = Gross Profit - Affiliate commission - Variable support cost

LTV = Monthly Contribution Margin per customer × Average retained months

CAC Payback = CAC ÷ Monthly Contribution Margin per customer
```

### 12.2. Example self-serve cohort

Assumptions:

- average monthly revenue: $35;
- payment and refund cost: 7%;
- AI and delivery COGS: 28%;
- affiliate and variable support: 8%;
- contribution margin: approximately $20 per month;
- average retention: 7 months.

Estimated LTV: approximately $140. A CAC above $70 would create a payback period above 3.5 months and leave limited room for error.

### 12.3. Break-even examples

If fixed operating burn is $20,000 per month and contribution margin per paying customer is $20, the business needs approximately 1,000 active paying customers to cover operating burn.

If blended contribution margin is $35, approximately 572 active paying customers are required.

These examples exclude founder compensation, taxes and one-time capital recovery unless explicitly included.

---

## 13. Ninety-day launch plan

### Days 1–30: validation and foundations

- select target segment;
- confirm brand and jurisdiction;
- validate provider availability and cost;
- build financial model by model and resolution;
- implement authentication, credit ledger and one payment path;
- launch three to five core tools;
- define trial and anti-abuse rules;
- instrument analytics;
- prepare legal documents;
- recruit design partners.

### Days 31–60: private beta

- onboard 50–150 target users;
- collect qualitative feedback;
- measure first successful generation;
- monitor COGS by user and model;
- fix failed-job and refund workflows;
- create case studies;
- test two pricing structures;
- begin creator seeding;
- establish support response process.

### Days 61–90: controlled public launch

- launch use-case pages;
- activate paid search and retargeting;
- publish creator content;
- launch affiliate tracking;
- introduce top-ups and subscription tests;
- remove high-abuse free access;
- review cohorts weekly;
- choose whether to scale, narrow or pause.

### Exit criteria after 90 days

- repeat usage from target segment;
- gross margin within target range;
- acceptable failure and refund rate;
- measurable paid conversion;
- positive retention trend;
- at least one repeatable acquisition channel;
- no unresolved critical billing or data-security issues.

---

## 14. Risk register

| Risk | Impact | Mitigation |
| :--- | :--- | :--- |
| Provider price increase | Margin compression | Dynamic price configuration and multi-provider routing |
| Model deprecation | Broken tools | Abstraction layer, versioning and replacement plan |
| Video abuse | High COGS | Strict trial limits, duration caps and fraud controls |
| Payment restrictions | Revenue interruption | More than one payment/MoR option |
| Tax non-compliance | Legal and financial exposure | Merchant of Record or professional tax support |
| Low retention | Paid growth becomes uneconomic | Segment focus, onboarding and reusable workflows |
| High refund/dispute rate | Payment-account risk | Clear pricing, status tracking and support process |
| Copyright or likeness claims | Legal exposure | Acceptable-use policy, reporting process and user consent requirements |
| Data incident | Trust and legal risk | Access controls, logging, minimisation and incident process |
| Currency exposure | Cost volatility | USD pricing, reserve margin and regular repricing |
| Founder bottleneck | Slow execution | Documentation, ownership and delegated operations |

---

## 15. Financial controls and reporting

Required weekly reporting:

- gross revenue;
- net revenue;
- subscription MRR;
- top-up revenue;
- COGS by provider and model;
- gross margin;
- payment fees;
- refunds and disputes;
- free-credit COGS;
- affiliate commission;
- paid acquisition spend;
- CAC by channel;
- active paying users;
- retention cohorts;
- outstanding provider liabilities.

Required controls:

- daily credit-ledger reconciliation;
- idempotent payment webhooks;
- provider usage reconciliation;
- model-level pricing configuration;
- spending alerts;
- refund approval rules;
- affiliate fraud review;
- monthly cash forecast;
- three-month minimum runway target.

---

## 16. Open decisions

- final target segment;
- launch jurisdiction and entity;
- Merchant of Record or direct processor;
- exact free-access structure;
- monthly plan prices and credit values;
- purchased-credit expiry policy;
- annual-plan timing;
- provider mix and failover strategy;
- whether video is available in the trial;
- initial paid-media ceiling;
- creator commission model;
- required runway before public launch;
- support coverage and SLA;
- data-retention periods;
- launch languages after English.

---

## Conclusion

The international platform can be launched with relatively low fixed infrastructure cost, but it is not a low-capital business once product development, AI COGS, customer acquisition and operational risk are included.

The financially strongest path is:

1. narrow the initial segment;
2. launch a small, metered product set;
3. avoid open-ended free generation;
4. validate retention before scaling paid media;
5. maintain model-level margins;
6. use a tax and payment structure appropriate for global sales;
7. keep enough runway for at least two full growth experiments.