Practice
SovTech
Open public infrastructure for sovereign finance.
Sovereign-debt analysis still depends heavily on opaque spreadsheets, fragmented data, and repetitive manual work. We build transparent, testable tools that carry more of that toil, so skilled professionals can spend their time on the analysis and judgment that matter.
Free to use. Open to inspect. Built to extend.
At the foundation of SovTech is free, open-source analytical infrastructure for sovereign debt. The core tools, methods, and standards we build with philanthropic support are published under permissive licenses so that ministries, international institutions, researchers, and advisors can inspect them, run them themselves, and build on them. The code lives at github.com/teal-insights.
Open licensing is a practical commitment, not a temporary free tier. Public releases can be self-hosted, modified, or maintained independently of Teal Insights. We aim to use open formats, documented methods, and clear exports so that adopting a tool does not trap an institution inside it.
Public infrastructure underneath. Expert work around it.
Open source does not mean that every engagement is free or every client project is public. Institutions can hire Teal Insights to adapt, integrate, deploy, maintain, or extend the public foundation. Paid work may include institutional deployment, private-data integration, custom modules, training, support, hosting, security, validation, and ongoing maintenance.
Clients pay for expert work and institution-specific value, not for permission to keep using a tool on which they have come to depend.
Why the model works
Philanthropic funding creates reusable infrastructure for the field. Open licensing reduces adoption risk and allows an ecosystem to grow. Paid institutional work supports the team that maintains the infrastructure and applies it to difficult real-world problems. A stronger public foundation makes Teal Insights more valuable; a durable Teal Insights makes the public foundation stronger.
Tools
Tool
Research preview · Open source
Project debt-to-GDP through 2031 for 170 economies under your own assumptions for growth, interest rates, the primary balance, and the exchange rate. A rebuild of the Financial Times' 2014 debt dynamics visualizer on the IMF's April 2026 World Economic Outlook, with the debt identity in the open and every default input carrying its source.
- Intended users
- Finance ministries, debt managers, researchers, advisors
- Status
- Research preview
- Partner
- NatureFinance
- License
- MIT
- Deployment
- Runs in the browser; a static build with no server
- Validation
- At reset, reproduces the WEO's own published debt-to-GDP path; each default input carries its origin: published, derived, or assumption
- Services
- Customization, integration, and methodology support available
Open the tool →View the source →
Tool
Research preview · Open source
Run the IMF's Quantitative Climate Risk Assessment Fiscal Tool (Q-CRAFT) without the spreadsheet: long-run projections of GDP, the fiscal balance, and debt under different climate scenarios for 175 countries through 2099. Cited in Uganda's Ministry of Finance interim report on embedding adaptation and resilience into the country's sovereign debt and credit profile.
- Intended users
- Finance ministries, fiscal risk units, researchers, advisors
- Status
- Research preview
- Partner
- NatureFinance
- Based on
- The IMF Fiscal Affairs Department's Q-CRAFT Excel tool
- License
- MIT
- Deployment
- Hosted web application; the open R source can also be run locally
- Validation
- Parity testing reproduces the IMF Excel workbook's outputs row for row on the Uganda country workbook (3,138 rows compared, zero deviation); the parity report ships with the source
- Services
- Institutional deployment, training, and methodology support available
Open the explorer →Read the companion guide →View the source →
Tool
Research preview · Open source
Examine a debt-for-development swap from three seats at the table: what the country saves, what the funder buys, and what the investor gives up. The model starts from actual cash flows and makes the assumptions explicit.
- Intended users
- Finance ministries, funders, investors, advisors
- Status
- Research preview
- License
- MIT
- Deployment
- Runs in the browser; a static page with no build step
- Validation
- Reproduces the World Bank Debt Swap Calculator's published worked example line by line; the methodology page documents scope and limits
- Services
- Customization, integration, and methodology support available
Open the explorer →Read the methodology →View the source →
Tool
Live · MIT licensed
Search and read 9,774 sovereign bond prospectuses and related filings from SEC EDGAR, the Luxembourg Stock Exchange, the FCA National Storage Mechanism, and Georgetown's #PublicDebtIsPublic corpus. Filter by country, region, and income group; search full document text; or fork the open-source application and run your own copy.
- Intended users
- Finance ministries, investors, researchers, legal advisors
- Status
- Live
- License
- MIT
- Deployment
- Browser-based; fork the open repository and run your own copy
- Validation
- Every document carries its public source; corpus and ingestion code are published
- Services
- Institutional deployment, private-data integration, and support available
Open the database →View the source →
Data
Dataset
v0.1.5 · CC BY 4.0
The foreign-currency share of public debt, read from the debt sustainability analyses published in IMF staff reports: 167 values found across the 191-country roster, with the 24 misses recorded as gaps. Every value carries the report, the PDF hash, the page, and the table or figure it came from, so any number can be checked against its source in about a minute, by a person or an agent.
- Intended users
- Finance ministries, IMF and World Bank staff, researchers, advisors
- Status
- Published release; corrections ship signed, versioned, and in the open
- License
- CC BY 4.0 for the data, MIT for the code
- Coverage
- 167 values across a 191-country roster, both DSA frameworks
- Validation
- Deterministic recompute on every release, a second AI model family following the published directions, seven signed corrections, and page-pinned provenance on every row. Two predeclared machine audits failed their own gates and are disclosed in full
Read about the dataset →View the source →
Teal Insights is an independent, founder-owned, for-profit firm. Our current philanthropic funding supports open-source infrastructure for the field. Over time, paid research, advisory, implementation, and institutional services will help sustain the team and the ecosystem around that infrastructure.
