Tech Cost Projections - Production Launch

Executive Summary

This document provides detailed cost projections for the Toto ecosystem as user base grows. Costs are broken down by service category and include scaling scenarios from 1,000 to 100,000+ active users.

Key Cost Drivers:

AI/ML Services (Google Gemini) - Largest variable cost
Firebase App Hosting (Cloud Run) - Scales with traffic
Firestore Database - Scales with data operations
Firebase Storage - Scales with image/file uploads
Email Services (SendGrid) - Scales with user engagement
Payment Processing (Stripe) - Transaction-based fees

🏗 Infrastructure Overview

Firebase Projects

toto-f9d2f - Main application (production)
toto-f9d2f-stg - Main application (staging)
toto-bo - Backoffice (production)
toto-bo-stg - Backoffice (staging)
toto-ai-hub - AI services
toto-wallet - Payment processing
toto-landing - Landing pages
toto-investors - Investor pages

Current Configuration

toto-app (Production):

Min instances: 2
Max instances: 100
CPU: 2 vCPU
Memory: 1024 MiB
Concurrency: 100

toto-bo (Production):

Min instances: 1
Max instances: 10
CPU: 1 vCPU
Memory: 1024 MiB
Concurrency: 80

toto-ai-hub:

Min instances: 2
Max instances: 20
CPU: 2 vCPU
Memory: 2048 MiB
Concurrency: 100

💵 Cost Breakdown by Service

1. AI/ML Services (Google Gemini) 🔴 HIGHEST COST

Current Usage:

Gemini 2.0 Flash (primary, 85-90% of usage) - $0.075 per 1M input tokens, $0.30 per 1M output tokens
Gemini 2.0 PRO (complex tasks, 10-15% of usage) - $1.25 per 1M input tokens, $5.00 per 1M output tokens
Model Selection Service optimizes Flash vs PRO usage (saves 30-50%)

Realistic Token Usage per Interaction:

System prompt: 800-1,200 tokens (agent persona, rules, context)
User message: 100-200 tokens
Knowledge context (RAG): 200-500 tokens
Conversation history: 200-500 tokens (if multi-turn)
Average input: 1,200-2,200 tokens per interaction
Average output: 200-400 tokens per interaction

Estimated Usage per User:

2-3 sessions per month
2-3 AI interactions per session
Total: 6-9 AI interactions per user per month
Plus: Intent detection, image analysis, RAG retrieval

Monthly Cost Projections (Realistic Estimates):

Users	AI Interactions/Month	Flash (85%)	PRO (15%)	Monthly Cost
1,000	10,000	8,500 × $0.000285	1,500 × $0.00575	$14-50
5,000	50,000	42,500 × $0.000285	7,500 × $0.00575	$70-250
10,000	100,000	85,000 × $0.000285	15,000 × $0.00575	$140-500
25,000	250,000	212,500 × $0.000285	37,500 × $0.00575	$350-1,250
50,000	500,000	425,000 × $0.000285	75,000 × $0.00575	$700-2,500
100,000	1,000,000	850,000 × $0.000285	150,000 × $0.00575	$1,400-5,000

Note: Cost range reflects conservative (higher) vs realistic (lower) estimates. Actual costs depend on:

Token usage per interaction (varies with context size)
PRO model usage percentage (varies with conversation complexity)
Additional services (embeddings, vector search if enabled)

Cost Optimization (Already Implemented):

✅ Model Selection Service (30-50% savings) - Automatically uses Flash for simple tasks
✅ Semantic Caching (reduces duplicate queries) - Saves 10-20%
✅ Function Calling (reduces token usage) - Saves 5-10%

Additional Optimization Opportunities:

⚠️ Rate limiting per user (potential 20-30% savings)
⚠️ Prompt optimization (potential 15-20% savings)
⚠️ Context window management (potential 10-15% savings)

Annual Cost Range: $168 - $60,000 (1K - 100K users, realistic estimates)

For detailed AI cost breakdown, see: AI_COSTS_DETAILED.md

2. Firebase App Hosting (Cloud Run) 🟡 MEDIUM-HIGH COST

Pricing:

CPU: $0.00002400 per vCPU-second
Memory: $0.00000250 per GiB-second
Requests: $0.40 per million requests
Minimum billable time: 100ms

toto-app Costs:

Users	Avg Requests/Month	Avg Instances	CPU Hours	Memory Hours	Monthly Cost
1,000	500K	2-5	720	360	$180
5,000	2.5M	3-8	1,800	900	$450
10,000	5M	5-15	3,600	1,800	$900
25,000	12.5M	8-25	9,000	4,500	$2,250
50,000	25M	15-50	18,000	9,000	$4,500
100,000	50M	25-100	36,000	18,000	$9,000

toto-bo Costs (Lower Traffic):

Users	Avg Requests/Month	Avg Instances	Monthly Cost
1,000	50K	1-2	$25
5,000	250K	1-3	$125
10,000	500K	1-5	$250
25,000	1.25M	2-8	$625
50,000	2.5M	3-10	$1,250
100,000	5M	5-10	$2,500

toto-ai-hub Costs:

Users	Avg Requests/Month	Avg Instances	Monthly Cost
1,000	6K	2-3	$50
5,000	30K	2-5	$250
10,000	60K	2-8	$500
25,000	150K	3-12	$1,250
50,000	300K	5-20	$2,500
100,000	600K	8-20	$5,000

Total App Hosting Monthly: $255 - $16,500 (1K - 100K users)

3. Firestore Database 🟡 MEDIUM COST

Pricing:

Document reads: $0.06 per 100K
Document writes: $0.18 per 100K
Document deletes: $0.02 per 100K
Storage: $0.18 per GiB/month
Network egress: $0.12 per GiB (first 10 GiB free)

Estimated Operations per User:

Reads: ~500/month (case views, user data, etc.)
Writes: ~50/month (updates, new data)
Storage: ~5MB/user (cases, images metadata, etc.)

Monthly Cost Projections:

Users	Reads/Month	Writes/Month	Storage (GB)	Monthly Cost
1,000	500K	50K	5	$330
5,000	2.5M	250K	25	$1,650
10,000	5M	500K	50	$3,300
25,000	12.5M	1.25M	125	$8,250
50,000	25M	2.5M	250	$16,500
100,000	50M	5M	500	$33,000

Cost Optimization:

✅ Index optimization (reduces unnecessary reads)
✅ Caching strategies
⚠️ Consider: Data archival for old cases

Annual Cost Range: $3,960 - $396,000 (1K - 100K users)

4. Firebase Storage 🟢 LOW-MEDIUM COST

Pricing:

Storage: $0.026 per GiB/month
Downloads: $0.12 per GiB (first 1 GiB free/day)
Uploads: $0.05 per GiB

Estimated Usage per User:

Average storage: 50MB/user (case images, profile photos)
Downloads: 200MB/user/month (image views)

Monthly Cost Projections:

Users	Storage (GB)	Downloads (GB)	Monthly Cost
1,000	50	200	$30
5,000	250	1,000	$150
10,000	500	2,000	$300
25,000	1,250	5,000	$750
50,000	2,500	10,000	$1,500
100,000	5,000	20,000	$3,000

Annual Cost Range: $360 - $36,000 (1K - 100K users)

5. Email Services (SendGrid) 🟢 LOW COST

Pricing:

Free tier: 100 emails/day
Essentials: $19.95/month for 50K emails
Pro: $89.95/month for 100K emails
Additional: $0.20 per 1,000 emails over plan

Estimated Usage:

Welcome emails: 1 per new user
Transactional: 2-3 per active user/month
Notifications: 1-2 per active user/month

Monthly Cost Projections:

Users	Emails/Month	Plan	Monthly Cost
1,000	3,000	Free	$0
5,000	15,000	Essentials	$20
10,000	30,000	Essentials	$20
25,000	75,000	Pro	$90
50,000	150,000	Pro + 50K	$190
100,000	300,000	Pro + 200K	$390

Annual Cost Range: $0 - $4,680 (1K - 100K users)

6. Payment Processing (Stripe) 🟡 TRANSACTION-BASED

Pricing:

Standard: 2.9% + $0.30 per transaction
International: 3.9% + $0.30
ACH: 0.8% (max $5)

Estimated Usage:

Average donation: $50
Conversion rate: 5% of active users donate monthly
Average transactions per donor: 1.5/month

Monthly Cost Projections:

Users	Donors	Transactions	Revenue	Stripe Fees
1,000	50	75	$3,750	$109
5,000	250	375	$18,750	$544
10,000	500	750	$37,500	$1,088
25,000	1,250	1,875	$93,750	$2,719
50,000	2,500	3,750	$187,500	$5,438
100,000	5,000	7,500	$375,000	$10,875

Note: These are fees, not costs - they're deducted from revenue.

Annual Fee Range: $1,308 - $130,500 (1K - 100K users)

7. Error Monitoring (Sentry) 🟢 LOW COST

Pricing:

Developer: Free (5K events/month)
Team: $26/month (50K events)
Business: $80/month (100K events)

Estimated Usage:

~10 events per user per month (errors, warnings)

Monthly Cost Projections:

Users	Events/Month	Plan	Monthly Cost
1,000	10K	Team	$26
5,000	50K	Team	$26
10,000	100K	Business	$80
25,000	250K	Business + Overage	$380
50,000	500K	Business + Overage	$880
100,000	1M	Business + Overage	$1,880

Annual Cost Range: $312 - $22,560 (1K - 100K users)

8. Twitter API 🟢 LOW COST (if used)

Pricing:

Basic: $100/month (10K tweets/month)
Pro: $5,000/month (1M tweets/month)

Estimated Usage:

Social media monitoring for guardians
~100 monitored accounts
~1,000 tweets analyzed/month

Monthly Cost: $100 (Basic plan sufficient for most scenarios)

Annual Cost: $1,200

9. Vertex AI Vector Search 🟡 OPTIONAL

Pricing:

Index creation: $0.10 per GB indexed
Query: $0.10 per 1,000 queries
Storage: $0.10 per GB/month

Current Status: Optional (code supports it, but using in-memory by default)

If Enabled:

Estimated: 1M vectors, 10K queries/month
Monthly Cost: ~$200

📊 Total Monthly Cost Summary

Note: AI costs shown as ranges (realistic - conservative estimates)

Users	AI (Range)	App Hosting	Firestore	Storage	Email	Sentry	Twitter	TOTAL (Range)
1,000	$14-50	$255	$330	$30	$0	$26	$100	$755-791
5,000	$70-250	$825	$1,650	$150	$20	$26	$100	$2,841-3,021
10,000	$140-500	$1,650	$3,300	$300	$20	$80	$100	$5,590-5,950
25,000	$350-1,250	$4,125	$8,250	$750	$90	$380	$100	$14,045-14,945
50,000	$700-2,500	$8,250	$16,500	$1,500	$190	$880	$100	$28,120-29,920
100,000	$1,400-5,000	$16,500	$33,000	$3,000	$390	$1,880	$100	$56,270-60,870

Annual Cost Range: $9,060 - $730,440 (1K - 100K users, realistic estimates)

Key Changes from Previous Estimates:

AI costs updated to reflect realistic token usage patterns
Lower end of range represents typical usage with optimizations
Higher end represents conservative estimates with buffers
See AI_COSTS_DETAILED.md for detailed breakdown

🎯 Cost Optimization Strategies

Immediate Actions (Already Implemented)

✅ Model Selection Service - Saves 30-50% on AI costs
✅ Function Calling - Reduces token usage
✅ Semantic Caching - Avoids duplicate AI queries

Recommended Actions

AI Cost Optimization:
- Implement rate limiting per user (e.g., 10 AI interactions/day free, then paid)
- Add usage quotas for premium features
- Cache common queries/responses
- Consider batch processing for non-real-time tasks
Firestore Optimization:
- Implement data archival for cases older than 1 year
- Use composite indexes efficiently
- Batch operations where possible
- Consider read replicas for heavy read workloads
App Hosting Optimization:
- Right-size min instances based on actual traffic patterns
- Implement CDN caching for static assets
- Optimize bundle sizes to reduce cold start times
- Use regional deployments to reduce latency
Storage Optimization:
- Implement image compression/optimization
- Use CDN for frequently accessed images
- Archive old images to cheaper storage tiers
- Implement lazy loading for images
Email Optimization:
- Batch notifications where possible
- Implement email preferences (reduce unnecessary emails)
- Use transactional email templates efficiently

📈 Growth Scenarios

Conservative Growth (Year 1)

Month 1-3: 1,000 users → $1,191/month
Month 4-6: 5,000 users → $5,021/month
Month 7-9: 10,000 users → $9,950/month
Month 10-12: 15,000 users → ~$14,000/month

Year 1 Total: ~$90,000

Moderate Growth (Year 1)

Month 1-2: 1,000 users → $1,191/month
Month 3-4: 5,000 users → $5,021/month
Month 5-6: 10,000 users → $9,950/month
Month 7-8: 25,000 users → $24,945/month
Month 9-12: 50,000 users → $49,920/month

Year 1 Total: ~$240,000

Aggressive Growth (Year 1)

Month 1: 1,000 users → $1,191/month
Month 2-3: 5,000 users → $5,021/month
Month 4-5: 10,000 users → $9,950/month
Month 6-7: 25,000 users → $24,945/month
Month 8-9: 50,000 users → $49,920/month
Month 10-12: 100,000 users → $101,870/month

Year 1 Total: ~$600,000

⚠️ Cost Risk Factors

High Risk:

AI Usage Spikes - Viral content could cause 10x AI usage
Storage Growth - Image-heavy cases could exceed estimates
Database Scaling - Complex queries could increase Firestore costs

Medium Risk:

Traffic Spikes - Marketing campaigns could spike App Hosting costs
Email Volume - Notification preferences could increase email costs

Mitigation:

Set up billing alerts at 50%, 75%, 90%, 100% of budget
Implement usage quotas and rate limiting
Monitor costs daily during launch period
Have cost optimization plan ready

📋 Monitoring & Alerts

Recommended Billing Alerts:

Daily Budget: Set at 1/30th of monthly budget
Weekly Budget: Set at 1/4th of monthly budget
Monthly Budget: Set based on user growth projections

Key Metrics to Track:

AI token usage (input/output split)
Firestore read/write operations
App Hosting instance hours
Storage growth rate
Email send volume
Error rate (Sentry events)

🔄 Review Schedule

Weekly: Review cost trends and anomalies
Monthly: Compare actual vs projected costs
Quarterly: Update projections based on actual growth
Annually: Comprehensive cost optimization review

📞 Support & Questions

For questions about cost projections:

Technical Issues: Review Firebase/Google Cloud billing console
Optimization Ideas: Review this document's optimization strategies
Budget Planning: Use growth scenarios as starting point

Last Updated: January 2025
Next Review: Quarterly
Status: 📊 Active Monitoring

Executive Summary​

🏗 Infrastructure Overview​

Firebase Projects​

Current Configuration​

💵 Cost Breakdown by Service​

1. AI/ML Services (Google Gemini) 🔴 HIGHEST COST​

2. Firebase App Hosting (Cloud Run) 🟡 MEDIUM-HIGH COST​

3. Firestore Database 🟡 MEDIUM COST​

4. Firebase Storage 🟢 LOW-MEDIUM COST​

5. Email Services (SendGrid) 🟢 LOW COST​

6. Payment Processing (Stripe) 🟡 TRANSACTION-BASED​

7. Error Monitoring (Sentry) 🟢 LOW COST​

8. Twitter API 🟢 LOW COST (if used)​

9. Vertex AI Vector Search 🟡 OPTIONAL​

📊 Total Monthly Cost Summary​

🎯 Cost Optimization Strategies​

Immediate Actions (Already Implemented)​

Recommended Actions​

📈 Growth Scenarios​

Conservative Growth (Year 1)​

Moderate Growth (Year 1)​

Aggressive Growth (Year 1)​

⚠️ Cost Risk Factors​

High Risk:​

Medium Risk:​

Mitigation:​

📋 Monitoring & Alerts​

Recommended Billing Alerts:​

Key Metrics to Track:​

🔄 Review Schedule​

📞 Support & Questions​

Executive Summary

🏗 Infrastructure Overview

Firebase Projects

Current Configuration

💵 Cost Breakdown by Service

1. AI/ML Services (Google Gemini) 🔴 HIGHEST COST

2. Firebase App Hosting (Cloud Run) 🟡 MEDIUM-HIGH COST

3. Firestore Database 🟡 MEDIUM COST

4. Firebase Storage 🟢 LOW-MEDIUM COST

5. Email Services (SendGrid) 🟢 LOW COST

6. Payment Processing (Stripe) 🟡 TRANSACTION-BASED

7. Error Monitoring (Sentry) 🟢 LOW COST

8. Twitter API 🟢 LOW COST (if used)

9. Vertex AI Vector Search 🟡 OPTIONAL

📊 Total Monthly Cost Summary

🎯 Cost Optimization Strategies

Immediate Actions (Already Implemented)

Recommended Actions

📈 Growth Scenarios

Conservative Growth (Year 1)

Moderate Growth (Year 1)

Aggressive Growth (Year 1)

⚠️ Cost Risk Factors

High Risk:

Medium Risk:

Mitigation:

📋 Monitoring & Alerts

Recommended Billing Alerts:

Key Metrics to Track:

🔄 Review Schedule

📞 Support & Questions