SEBIT-MCP Models (English) 🌍

한국어 설명은 아래 링크에서 확인할 수 있습니다.
👉 README.ko.md

SEBIT (Systematic Engineered Binancial Intelligence & Tactics) is an MCP-based framework specialized in Accounting & Finance.
It consists of 12 core models, each operating on structured JSON inputs and calculation logic.

🚀 Installation & Run

git clone https://github.com/sebit-provider/sebit-mcp-public.git
cd sebit-mcp-public
npm install sebit-mcp-public
node dist/mcp-server.js

Claude & Other Client Integration

This framework is MCP-based. To integrate with Claude Desktop (or other MCP clients), edit the claude_desktop_config.json file:

{
  "mcpServers": {
    "sebit-mcp": {
      "command": "C:\\Program Files\\nodejs\\node.exe",
      "args": ["C:\\Users\\user\\sebit-mcp-public\\dist\\mcp-server.js"],
      "cwd": "C:\\Users\\user\\sebit-mcp-public",
      "optional": true
    }
  }
}

📊 Model Descriptions

1. DDA (Dynamic Depreciation Algorithm)

Description: Calculates depreciation dynamically, factoring in time, usage, and market sensitivity.

🔹 Input example

{
  "acquisitionCost": 1200000,
  "residualValue": 200000,
  "usefulLifeYears": 5,
  "elapsedUseDays": 730,
  "periodUseDays": 180,
  "baselineUseHours": 2000,
  "totalUseHours": 2300,
  "beta": 0.4,
  "psPrev": 115,
  "psCurr": 108,
  "marketChangeR": 0.03
}

• Daily depreciation based on elapsed days
• Adjustments for over/under usage
• Market rate (r) and β sensitivity applied
• Impairment test & revaluation (cap/thresholds)

Relevant IFRS: IAS 16, IAS 36

2. LAM (Lease Asset Model)

Description: Evaluates lease liabilities and right-of-use (ROU) assets considering cost, rate, and usage.

🔹 Input example

{
  "acquisitionCost": 8000000,
  "residualValue": 300000,
  "leaseTermYears": 4,
  "daysUsedThisPeriod": 120,
  "totalDays": 1460,
  "discountRate": 0.055
}

• PV-based lease liability valuation
• Depreciation based on usage days
• Adjustment for residual value & unused term

Relevant IFRS: IFRS 16.23–35

3. RVM (Resource Valuation Model)

Description: Values resources using cumulative and current mining data with market adjustments.

🔹 Input example

{
  "cumulativeMiningDays": 1500,
  "cumulativeMinedValue": 75000,
  "currentPeriodMiningDays": 90,
  "currentPeriodMinedValue": 5600,
  "currentResourcePrice": 52,
  "prevYearValuation": 68000,
  "currentValuation": 73000
}

• Cumulative & period-based resource valuation
• Market price variation (r) and β sensitivity applied
• Comparative analysis with previous year

Relevant IFRS: IFRS 6, IAS 16

4. CEEM (Consumable Expense Estimation Model)

Description: Estimates consumable expenses based on cumulative usage and unit costs.

🔹 Input example

{
  "cumulativeUsage": 15000,
  "unitCost": 18.5,
  "periodDays": 90,
  "totalUsage": 1400,
  "prevYearR": 0.06,
  "beta": 0.8,
  "years": 2
}

• Expense = Unit cost × Usage
• Growth rate (r) and β sensitivity applied
• Aggregated per-period cost analysis

Relevant IFRS: IAS 2, IAS 16

5. BDM (Bond Discounting Model)

Description: Discounts bonds to present value considering issue amount, elapsed days, and discount rate.

🔹 Input example

{
  "issueAmount": 50000000,
  "scheduleDays": 1825,
  "elapsedDays": 365,
  "prevMeasuredValue": 48200000,
  "discountRate": 0.047
}

• Present Value (PV) based on elapsed days
• Adjustments with discount rate & β
• Comparison with prior valuations

Relevant IFRS: IFRS 9

6. BELM (Bad Debt Expected Loss Model)

Description: Estimates Expected Loss Rate (ELR) using settlements, exposures, and interest rates.

🔹 Input example

{
  "dailyExpectedSettlement": 35000,
  "usefulLifeYears": 6,
  "elapsedDays": 450,
  "actualSettlementToDate": 9200000,
  "interestRate": 0.065,
  "clientExposure": 60000000,
  "totalExposure": 1500000000
}

• Expected vs actual settlements
• Portfolio weighting and historical performance
• Final ELR (0–1 range)

Relevant IFRS: IFRS 9

7. CPRM (Convertible Bond Risk Model)

Description: Calculates convertible bond risk based on base rate, bad debts, PD, volumes, and recoveries.

🔹 Input example

{
  "baseRate": 0.05,
  "badDebtIncidence": 0.02,
  "assumedDefaultRate": 0.03,
  "bondUnitPrice": 1000,
  "bondVolume": 60000,
  "pastDebtorRecovery": 15000,
  "bondTurnoverPct": 0.55,
  "stockTurnoverPct": 0.72,
  "extraAdj": -0.004,
  "maxValue": 0.30
}

• Base rate + PD + Bad debt incidence
• Adjustments with trading volumes & recoveries
• Risk cap applied

Relevant IFRS: IFRS 9, IAS 32

8. OCIM (Other Comprehensive Income Model)

Description: Compounds OCI considering account shares, flows, sensitivity, and adjustments.

🔹 Input example

{
  "accountOCIAmount": 18000000,
  "totalOCIAllItems": 92000000,
  "openingOCIBalance": 50000000,
  "currentPeriodOCI": 13500000,
  "marketChangeR": 0.045,
  "beta": 1.1,
  "horizonYears": 4
}

• OCI account share calculation
• Compound evaluation of opening & current OCI
• Sensitivity and adjustment applied

Relevant IFRS: IFRS 9, IAS 1

9. FAREX (Foreign Exchange Adjustment Model)

Description: Adjusts FX based on export/import data and computes effective exchange rate.

🔹 Input example

{
  "prevYear_export_curr": 142000000,
  "prevYear_import_curr": 108000000,
  "currYear_export_curr": 160000000,
  "currYear_import_curr": 120000000,
  "currentExchangeRate": 1332
}

• Trade balance analysis (YoY comparison)
• FX sensitivity (β, weights) applied
• Effective exchange rate computed

Relevant IFRS: IAS 21

10. TCT-BEAM (Trigonometric Cost Tracking & BE Analysis Model)

Description: Uses trigonometric angles of fixed/variable costs to analyze revenue sensitivity and break-even.

🔹 Input example

{
    "fixedCosts": [920000000, 980000000, 1050000000, 1120000000,
  1200000000],
    "variableCosts": [450000000, 480000000, 520000000, 560000000,
  600000000],
    "currentRevenue": 2100000000,
    "options": {
      "language": "en",
      "includeGraph": true,
      "roundStep": 1000
    }
  }

• Conversion of costs into angular representation
• Break-even point (BEP) estimation
• Sensitivity analysis

Relevant IFRS: IAS 2, IAS 1

11. CPMRV (Crypto Market Real Value)

Description: Evaluates cryptocurrency fair value using past growth/decline rates and current market value.

🔹 Input example

{
  "previousYearGrowthRate": 0.42,
  "previousYearDeclineRate": 0.10,
  "currentYearGrowthYTD": 0.25,
  "currentYearDeclineYTD": 0.07,
  "currentCryptocurrencyValue": 48000,
  "horizonMonths": 24
}

• Historical growth/decline rates applied
• YTD adjustments
• Real (fair) value computed

Relevant IFRS: IAS 38, IAS 2, IFRS 13

12. DCBPRA (Dynamic CAPM-Based Pricing Risk Adjustment)

Description: Adjusts CAPM pricing with real growth rates to dynamically evaluate risk premium.

🔹 Input example

{
  "riskFreeRate": 0.025,
  "marketReturn": 0.082,
  "beta": 1.38,
  "RS": 0.15,
  "realGrowthPct": 0.039
}

• CAPM-based expected return
• RS & real growth adjustment
• Final risk-adjusted return

Relevant IFRS: IFRS 13, IAS 36, IAS 19

Additional Features(Since 1.0.6)

JOURNAL (Dual-language Journal Builder)

Description: Generates and maintains accounting journals in Excel format with Korean (분개장) and English (journal) ledgers.

🔹 Input example

{
  "company": "SEBIT Corp",
  "text": "2025-03-20 LG Electronics laptop purchase 2,500,000 KRW paid by bank transfer",
  "options": {
    "baseDir": "Desktop/journal_book",
    "oneWorkbookPerYear": true
  }
}

Natural Language Journal Entry(Updated at v1.0.8): You can now generate accounting journal entries without writing JSON. Just type a natural language sentence in English or Korean, and the system will parse and classify automatically.

I purchase LG Electronics laptop 2,500,000 KRW, paid by bank transfer.

🔹 Output example output filename: 2025_journal.xlsx Monthly sheets: 01 … 12 Audit log: audit.log

Features: ✅ Natural language → Journal entry (ko/en) ✅ Automatic account classification (API + heuristics) ✅ Duplicate check & audit logging ✅ Monthly sheets + SUMMARY sheet auto-updated

Relevant IFRS: IAS 1, IAS 2, IAS 16, IFRS 9

TCT-BEAM Trigonometric Graphs(Added 1.0.7)

Visualizes fixed and variable costs as trigonometric vectors. Provides a break-even chart with angle sensitivity (θ) and revenue-performance visualization. Output: SVG/PNG charts auto-generated for each run.

🔹 Example usage

{
  "fixedCosts": [850000000, 920000000, 995000000],
  "variableCosts": [420000000, 445000000, 485000000],
  "currentRevenue": 1850000000,
  "options": { "chart": true, "outputDir": "./reports" }
}

Output: beam_graph.svg

Automated Report Generation(Added 1.0.7)

Generates a structured PDF report for each MCP session.

Includes: ✅Model execution logs ✅Aggregated risk classification (Low / Medium / High) ✅Strategic roadmap (24h, 1 week, 1 month) ✅IFRS references

🔹 Example output SEBIT-MCP-Report_2025-09-17_17-03-50.pdf

📌 Notes

• All inputs must be in JSON format.
• Numeric fields allow string input ("8%", "0.08")
• Optional fields (options) may be added.
• See SEBIT_FRAMEWORK_INPUT_VALUABLES.docx for detailed input docs.
• All models comply with IFRS standards.

License Notice
This project is licensed under the Sebit Public License v1.0 (SPL-1.0).

• ✅ Free for personal, educational, and research purposes
• 💼 Commercial use requires a separate license from the Author
• ✍️ Attribution ("SEBIT") is required in any use or derivative work

For details, see the LICENSE and NOTICE files.

🧾 License & Author

• License: Sebit Public License v1.0(SPL-1.0)
• Author: Seounghyup Park (박승협)

📬 문의 (Contact)

• Email: [email protected]
• GitHub Issues: sebit-mcp-public Issues