Goal: 3 working, magical Gene Keys audio tracks delivered to real humans. Validate the core product before spending anything on ads or app builds.
| Tool |
Type |
What You Use It For |
Cost |
| Claude Proclaude.ai — you're already using it |
Paid |
Research agent, script writing, affirmation generation, prompt development, this whole project co-pilot |
$20/moalready have this |
| Anthropic APIClaude Sonnet 4 via API |
Usage |
Powers your actual production agents. Script Agent, QA Agent, Research Agent calls. Claude Sonnet 4 is ~$3 per million input tokens / $15 out. |
~$3–8for full Sprint 1 batch |
| ElevenLabsStarter plan |
Paid |
Voice rendering for all audio tracks. Starter gives 30k chars/mo (~3–4 full-length affirmation tracks). Creator plan ($22/mo) gives 100k chars — recommended. |
$22/moCreator plan |
| Swiss Ephemerispython-flatlib or kerykeion |
Free |
The gold-standard chart calculation library. Free, open-source, runs locally on Mac Mini. Calculates planetary degrees, Gene Keys mapping, HD bodygraph data. |
$0open source |
| NotebookLMGoogle — free tier |
Free |
Build your Gene Keys knowledge base. Upload YouTube transcripts, public domain I Ching texts, Gene Keys summary articles. Research agent's source of truth. |
$0free tier sufficient |
| Suno AIPro plan for commercial use |
Paid |
Generate custom music beds — celestial, ambient, solfeggio-inspired. Pro plan gives commercial license. Create 5–8 template tracks in Sprint 1. You own them. |
$10/moPro plan · 1 month |
| FFmpegaudio mixing CLI tool |
Free |
Mix voice + music layers. Adjust volumes, fade in/out, export final MP3. Runs on Mac Mini. Claude Code can write the scripts for you. |
$0open source |
| Claude CodeCLI — uses Anthropic API |
Usage |
Your primary dev tool on Mac Mini. Writes the Python agent scripts, Swiss Ephemeris integration, FFmpeg pipelines. Costs are API token usage — already counted above. |
~$0 extraincluded in API costs |
| Pinecone / ChromaDBvector database for RAG |
Free |
Store Gene Keys knowledge embeddings. ChromaDB runs entirely free and local on Mac Mini for Sprint 1. Pinecone free tier works if you prefer cloud. |
$0free local or cloud tier |
| Python environmentMac Mini setup |
You Own It |
Python 3.11+, pip packages (kerykeion, anthropic SDK, chromadb, elevenlabs). All free. Claude Code sets this up with you. |
$0 |
| Sprint 1 Total (one-time + first month) |
~$55–75 |
Sprint 1 reality check: Your first 3 demo tracks will cost roughly $55–75 total — mostly ElevenLabs + one month of Suno. The AI API costs for generating scripts are under $10 for the entire sprint. This is not a financial barrier. It's a time barrier.
Smart choice using GHL (GoHighLevel). If you already have a GHL subscription for your coaching business, you can build the entire Affirmology funnel, intake form, and member delivery area inside it at no extra cost. One subscription, two businesses.
| Tool |
Type |
What You Use It For |
Cost |
| GoHighLevelalready have for coaching |
You Own It |
Landing page, opt-in form, birth data intake form, member portal, audio file delivery, email sequences, Stripe payment integration. Zero extra cost if already subscribed. |
$0 extraalready paying ~$97/mo |
| Stripevia GHL integration |
Usage |
Payment processing. 2.9% + $0.30 per transaction. On a $67 sale you keep ~$64.75. No monthly fee. |
2.9% + $0.30per transaction only |
| Domain / SSLaffirmology.com or similar |
Paid |
Brand domain. Check: affirmology.com, affirmologyai.com, myaffirmology.com. GHL hosts pages, just need the domain pointed over. |
$12–15/yrone-time annual |
| MidjourneyBasic plan |
Paid |
Landing page hero visuals, ad creatives, cosmic imagery. One month to create an image library, then cancel or downgrade. |
$10/moBasic plan · 1–2 months |
| Canva Proif not already subscribed |
Paid |
Ad graphics, landing page design elements, social posts, email headers. Optional — GHL has basic design built in. |
$15/mooptional if needed |
| Loomfree tier |
Free |
Record a personal video for the landing page. A 60-second authentic founder video on a landing page often converts better than polished production. |
$0 |
| Phase 2 Added Cost (first month) |
~$25–40 |
| Tool / Channel |
Type |
Notes |
Cost |
| Meta AdsFacebook + Instagram |
Ad Spend |
Nano-test phase. Run 3–5 ad variants at $5–10/day each. Astrology, Gene Keys, Abraham Hicks audience targeting. Kill losers in 3 days, scale winner. |
$150–300first test month |
| CapCutfree mobile app |
Free |
Edit short-form video ads. Waveform overlays on audio clips look stunning and perform well. No design skill needed. |
$0 |
| AdCreative.aioptional |
Optional |
AI-generated ad creative variations. Useful later for scaling. Skip in Sprint 1 — do it manually with Canva first. |
Skip for now~$29/mo when ready |
| Free sample strategyinfluencer seeding |
Free |
Give 5–10 free personalized audios to Gene Keys practitioners, HD coaches, astrology micro-influencers. One authentic share in the right Facebook group can outperform $500 in ads. |
~$3–6 API costper audio seeded |
| Phase 3 Ad Budget (first test month) |
$150–300 |
| Tool |
Notes |
Monthly Cost |
| Claude Pro |
Your research + co-pilot. Already paying. |
$20 |
| Anthropic APIproduction agents |
Scales with users. 0–50 users: ~$10–30/mo. Each audio script ≈ $0.05–0.15 in tokens. |
$10–30 |
| ElevenLabs Creator |
100k chars/mo ≈ 10–12 full audio tracks. Scale up plan as user count grows. |
$22 |
| GoHighLevelshared with coaching biz |
Landing page, funnel, delivery, email, CRM. Already paying — zero marginal cost. |
$0 extra |
| Suno / Musicafter template library built |
Once you have 6–8 music beds created, cancel Suno. Re-subscribe when you need new tracks. |
$0cancel after Sprint 1 |
| Domain + hosting |
Domain is annual. GHL handles hosting — no Vercel/Netlify cost needed. |
~$1 |
| Epidemic Soundoptional music alternative |
Pre-cleared music library. Good alternative to Suno if you want human-made tracks. Personal plan $10/mo — commercial requires Creator at $15/mo. |
$0–15optional |
| Monthly Recurring (early stage, no ads) |
$53–73/mo |
The beautiful math: At $53–73/mo overhead, you break even with just 1 Starter Kit sale per month (at $67). Every sale after that is profit. The cost structure is tiny relative to the price point.
| Step |
Tool |
Approx Cost Per Audio |
| Chart calculation |
Swiss Ephemeris (local) |
$0.00 |
| RAG knowledge lookup |
ChromaDB (local) |
$0.00 |
| Script generation (~3k tokens in, ~1.5k out) |
Claude Sonnet 4 API |
$0.03–0.08 |
| QA agent check (~1k tokens) |
Claude Sonnet 4 API |
$0.02–0.04 |
| Voice render (~5,000–8,000 chars for 10min audio) |
ElevenLabs Creator |
$0.08–0.15 |
| Music bed (reused template, pre-made) |
Suno (already paid, reused) |
$0.00 |
| Audio mixing + export |
FFmpeg (local) |
$0.00 |
| Total Cost Per Audio |
$0.13–0.27 |
Margin math: At $67 Starter Kit (includes 3–5 audios), your production cost is roughly $0.65–$1.35 total. 98%+ gross margin on the audio product itself. The overhead is fixed subscriptions, not variable production costs.
✓ Ways to Reduce / Avoid It
- PWA (Progressive Web App) — a website that feels like an app. No App Store. No 30%. Works on iOS and Android. This is the smart Sprint 3 move.
- Web-only checkout — Apple's rules say the 30% applies to in-app purchases, not web purchases. If users buy on your website and log in via the app, you keep 100%.
- External payment links (iOS 17.4+) — Apple now allows US apps to link to external payment pages with a small warning dialog. You pay Apple a 27% fee instead of 30% — marginal savings but it's something.
- Android / Google Play — Google allows external billing with ~4% reduced fee (26% instead of 30%). Better than Apple.
- Subscription threshold — Apple drops their cut to 15% after a subscriber has been with you for 12 months. Long-term subscribers get cheaper.
$ Does It Make Financial Sense?
- At $19/mo membership, Apple takes $5.70/mo per subscriber via in-app purchase. PWA avoids this entirely.
- At 100 subscribers, App Store costs you $570/mo vs $0 on PWA. That's real money.
- PWA recommendation: Build as a PWA first. 90% of the user experience is identical. Avoid the store entirely until you have evidence users strongly prefer native.
- Native app is worth it eventually — push notifications convert 3–5x better for monthly transit drops. But it's a Phase 4 problem, not now.
- Never try to hide in-app purchase options to avoid Apple's cut — this violates TOS and gets you kicked. The PWA route is the ethical, smart path.
Sprint 1 Demo
One-Time Setup
$55–75
ElevenLabs + Suno month + API costs. Everything else is free.
Phase 2 Launch
Landing Page
+$25–40
Domain + Midjourney month. GHL already paid for coaching.
Phase 3 Ads
First Test Month
+$150–300
Nano-test budget. Kill losers fast. 2 sales covers this.
Monthly Overhead
Ongoing
$53–73
Claude Pro + API + ElevenLabs. 1 sale/mo covers it.
Break-Even Point
2–3 sales
At $67 Starter Kit. Then every dollar is profit + reinvestment.
Cost Per Audio
Produced
$0.13–0.27
98%+ gross margin. The economics are exceptional.
Bottom line: You can get from zero to 3 real demos in people's hands for under $75. A working funnel with first ad tests for under $400 total. This is one of the most capital-efficient AI businesses possible because your COGS (cost of goods sold) is essentially API tokens and voice characters — both of which cost pennies and scale gracefully.
Claude Code tip: Start with the free tier of the Anthropic API — you get $5 in free credits to begin. Claude Code itself is free to install; it just uses your API balance. Your Mac Mini is the perfect local dev environment: run Swiss Ephemeris calculations, ChromaDB vector store, FFmpeg mixing, and Claude Code orchestration all for $0 in infrastructure. The cloud only gets involved when you call ElevenLabs or the Anthropic API.