Sigillum Finans documentation
Guides for integrating services, reviewing AI transaction traces and verifying recorded evidence.
#Purpose and scope
Sigillum Finans links observable steps in your institution’s AI-enabled applications. Review inputs, model and tool calls, source references, application-provided explanations and outcomes within the context of a transaction.
Integrity verification checks recorded content against its cryptographic seals. Comparison with an institution’s own records checks the fields selected in a comparison definition. Trace review, integrity verification and source comparison answer different questions.
#Choose a starting point
| Task | Guide |
|---|---|
| Connect your first service | Quickstart |
| Configure the portal | Institution setup |
| Add application context | Additional observations |
| Verify recorded evidence | Integrity verification |
| Compare with a source record | Comparison workflow |
#Interpret observations correctly
Sigillum records observable information; it does not reconstruct a model’s hidden reasoning. An explanation supplied by the source application can be recorded. A verified integrity result does not certify a lending decision, model fairness or regulatory compliance.
The product runs within the institution’s environment. This public documentation website is not an institutional installation and does not accept transaction data, service credentials or evidence packages.
#Documentation version
This documentation is based on Sigillum Finans 0.44.0 and the .NET, Java and Python SDKs 0.11.0. Use the SDK package and matching configuration supplied by your institution’s portal. All downloadable examples use synthetic data.