The CemaAI ecosystem consists of 10 specialized AI agents coordinated through a central orchestration layer, connected via MCP (Model Context Protocol) servers for seamless integration with external services.
Sign in to like and favorite skills
The CemaAI ecosystem consists of 10 specialized AI agents coordinated through a central orchestration layer, connected via MCP (Model Context Protocol) servers for seamless integration with external services.
User Input → API Gateway → Auth Check → Credit Deduction (1-15 credits) ↓ SmartBooks AI receives input ↓ [Tools Activated] 1. OCR Engine (if image) → Extract text from receipt 2. Currency Converter → Handle multi-currency 3. Category Classifier (AI) → Auto-categorize expense 4. Tax Calculator → Compute tax implications 5. Anomaly Detector → Flag unusual transactions ↓ Data stored in DB_FIN (Supabase) ↓ [MCP Integrations Triggered] - Google Sheets MCP → Sync to user's spreadsheet - Calendar MCP → Set payment reminders - Email MCP → Send weekly financial summary - Payment Gateway MCP → Process subscription ↓ Output: Dashboard update, notifications sent
User Input → Auth → Credit Deduction (1-15 credits) ↓ CemaLegal AI analyzes request ↓ [Tools Activated] 1. Template Selector → Choose appropriate legal template 2. Clause Generator (AI) → Create custom clauses 3. Risk Analyzer → Identify unfavorable terms 4. Plain Language Translator → Explain legal jargon 5. Compliance Checker → Ensure jurisdiction compliance ↓ [LLM Integration] GPT-4/Claude generates contract with legal precision ↓ [MCP Integrations] - DocVault AI → Store generated contract - Google Drive MCP → Backup to user's Drive - Email MCP → Send to all parties - E-signature API → Enable signing ↓ Output: Contract ready for review/signing
User Query → Auth → Credit Deduction (1-10 credits) ↓ MyAI analyzes query complexity ↓ [Tools Activated] 1. Web Search Engine → Find relevant sources 2. Fact Checker → Verify claims against multiple sources 3. Summarizer (AI) → Condense long articles 4. Citation Generator → Track source attribution 5. Bias Detector → Identify potential bias in sources ↓ [LLM Integration] GPT-4 synthesizes information from multiple sources ↓ [Vector Database] Embedding Models → Store for future reference ↓ Output: Comprehensive answer with citations
User Creates Post → Auth → Credit Check (0-5 credits) ↓ CemaTea AI receives content ↓ [Pre-Publish Tools] 1. Content Moderator (AI) → Check for violations - Hate speech detection - Violence/graphic content - Spam detection - NSFW classification 2. Sentiment Analyzer → Gauge emotional tone 3. Language Detector → Identify language(s) 4. Hashtag Suggester → Recommend trending tags ↓ [If flagged] → Moderation Queue (human review) [If approved] → Publish to feed ↓ [Post-Publish Processing] 5. Feed Algorithm → Determine visibility - User interests - Engagement history - Recency score - Trending factor 6. Notification Engine → Alert followers 7. Engagement Tracker → Count views, likes, shares 8. Earnings Calculator → Compute creator rewards ↓ [MCP Integrations] - SMS/WhatsApp MCP → Notify followers - Analytics Engine → Track trends - Payment MCP → Process creator payouts ↓ Output: Post visible in feeds, analytics updated
CemaTea Post mentions scam → Auto-trigger CemaSafe AI ↓ "Did you mean to report this scammer? [Report Now]" CemaTea Business complaint post → Suggest CemaComplaint AI ↓ "File official complaint for better resolution"
New Learner → Onboarding Assessment ↓ CemaLearn AI creates diagnostic quiz (5-10 questions) ↓ [Assessment Tools] 1. Knowledge Gap Analyzer → Identify what user knows/doesn't know 2. Learning Style Detector → Visual/Auditory/Kinesthetic preference 3. Pace Evaluator → Determine learning speed ↓ [Path Generation] AI generates personalized curriculum ↓ Example: "Python for Data Science" path Module 1: Python Basics (5 lessons, 2 quizzes) Module 2: Data Structures (8 lessons, 3 quizzes, 1 project) Module 3: Pandas Library (10 lessons, 2 projects) ... Module 12: Capstone Project ↓ [During Lessons] User engages with content ↓ [Active Tools] 1. Comprehension Monitor → Ask check-in questions 2. Difficulty Adjuster → Make easier/harder in real-time 3. Note Generator (AI) → Auto-create study notes 4. Quiz Creator → Generate practice problems 5. Translator → Convert to user's language 6. Voice Engine → Read lessons aloud ↓ [Assessment Phase] User takes quiz/submits assignment ↓ [Grading Tools] 1. Auto-Grader (AI) → Score objective questions 2. Essay Analyzer → Grade written responses 3. Code Checker → Run and test code submissions 4. Feedback Generator → Provide detailed explanations ↓ [Adaptive Response] If score ≥ 90%: Unlock next module, increase difficulty If score 70-89%: Proceed with current difficulty If score 50-69%: Provide extra practice If score < 50%: Reteach with different approach ↓ [MCP Integrations] - Google Drive MCP → Store notes and assignments - Calendar MCP → Schedule study sessions - Email MCP → Progress reports to learner/parent - Certificate Generator → Issue completion certificates ↓ Output: Continuous learning journey
CemaLearn generates notes → Auto-save to DocVault AI ↓ Learner can access encrypted notes anytime CemaLearn Certificate → Share on CemaTea AI ↓ "I just completed Python Mastery! 🎓"
User Requests Credit Score → Auth → Credit Deduction (2-15 credits) ↓ CemaCredit AI gathers data from multiple sources ↓ [Data Collection via MCP] 1. SmartBooks API → Pull business revenue/expenses (12 months) 2. Njangi System → Contribution consistency 3. CemaPay API → Payment patterns, wallet balance 4. Bank API MCP → Account verification, balance trends 5. User Input → Self-reported debts, income ↓ [Score Calculation Engine] Factor 1: Payment History (35%) - On-time bill payments - Njangi completion rate - Late payment frequency Factor 2: Credit Utilization (30%) - Debt-to-income ratio - Outstanding balances - Available credit usage Factor 3: Savings Behavior (15%) - Regular deposits - Emergency fund size - Savings growth rate Factor 4: Financial Stability (10%) - Income consistency - Business longevity - Transaction volume Factor 5: Credit Age (10%) - Account age - Financial activity diversity ↓ [AI Analysis] LLM processes all factors → Generates 300-850 score ↓ [Recommendation Engine] AI identifies improvement opportunities ↓ Example: "Pay down ₦50K debt → +40 point increase" ↓ [Report Generation] 1. Basic Report: Score + factors (PDF) 2. Detailed Report: Full breakdown + trends (PDF) 3. Lender-Ready: + Lending recommendations (PDF) ↓ [MCP Integrations] - Email MCP → Send report - DocVault AI → Store securely - SMS MCP → Monthly score updates - Bank API MCP → Verify for lenders ↓ Output: Credit score, actionable insights, reports
SmartBooks records payment → CemaCredit updates score ↓ "Your on-time payment improved your credit score by 2 points!" CemaCredit shows low score → Suggest CemaLearn ↓ "Take our Financial Literacy course to improve money management"
User Checks Phone Number → Auth → FREE (0 credits) ↓ CemaSafe AI queries database ↓ [Number Verification] 1. Database Lookup → Check existing reports 2. Pattern Matcher → Compare to known scam patterns 3. Risk Scorer → Calculate 0-100 risk score ↓ Response within 2 seconds: "⚠️ DANGER: 47 reports, 92/100 risk score" ↓ --- OR --- ↓ User Submits Report (Scam/Harassment/Corruption) ↓ [Report Processing] 1. Anonymizer → Strip identifying info (if anonymous selected) 2. Evidence Encryptor → Secure all attachments 3. Categorizer (AI) → Classify report type 4. Sentiment Analyzer → Detect urgency/trauma 5. Pattern Detector → Cross-reference similar reports ↓ [If Multiple Similar Reports Detected] "ALERT: 5 other reports match this pattern" ↓ [Safety Resources] AI provides immediate support resources - Hotlines - Legal aid - Counseling - Shelters ↓ [Community Protection] 6. Alert Generator → Create community warning 7. Number Blacklist → Auto-block verified scammers 8. Analytics Engine → Track trends by region ↓ [MCP Integrations] - SMS/WhatsApp MCP → Send community alerts - Email MCP → Send support resources - Government API MCP → Report to authorities (with permission) ↓ [Admin Notification] If critical (violence, child safety) → Immediate escalation ↓ Output: Report recorded, resources provided, community protected
CemaTea user posts about scam → CemaSafe detects keywords ↓ "Would you like to report this officially? It helps protect others." CemaSafe report filed → Angel AI provides follow-up support ↓ "We see you filed a safety report. Need someone to talk to?"
User Contacts Support → Auth → FREE (standard support) ↓ Angel AI receives message ↓ [Initial Analysis] 1. Intent Classifier → What does user need? - Technical issue - Billing question - Refund request - Feature inquiry - Complaint 2. Sentiment Analyzer → How frustrated are they? - Happy (standard response) - Neutral (standard response) - Frustrated (empathetic response) - Angry (priority + empathy) 3. Urgency Detector → How critical is this? ↓ [Ticket Creation] Ticket #AS-2025-XXXXX generated Priority assigned: Low/Medium/High/Urgent ↓ [Knowledge Base Search] AI searches previous solutions ↓ [If Solution Found] → Provide immediate answer [If Uncertain] → Ask clarifying questions [If Complex] → Escalate to human ↓ --- REFUND REQUEST PATH --- ↓ [Refund Eligibility Check] 1. Transaction Verifier → Pull transaction details 2. Usage Calculator → How much service was used? 3. Policy Checker → Does it qualify? 4. Timeline Validator → Within refund window? ↓ [Decision Matrix] Eligible: Auto-approve, process immediately Partial: Offer prorated refund Ineligible: Explain + offer alternatives