Portable models
The trac model runtime
We understand the pain involved in building, deploying and using models in a variety of different contexts. That’s why we built the trac model runtime. A lightweight and open-source package for building portable, production-strength Python models that can be executed not only in the trac platform but also in your IDE, a data science notebook or any other production system.
Explore the model runtime The trac guarantee
Comprehensive controls built in at the platform level
Auditable processes
A complete version history of every object and action.
Repeatable calculations
Any calculation can be replicated in full with just three clicks.
Zero change risk
Self-serve with confidence — trac guarantees zero change risk.
Self-documenting
Fully-automated governance documentation
A complete metadata representation of all the platform objects, processes, and calculations ensures you have comprehensive, governance-ready documentation at the your fingertips, with no manual input needed.
Governance ready
Audit, model governance and data lineage reporting
An append-only data model ensures that a complete, immutable version history is retained for every data object. The automated governance report for data objects includes:
- Object & version ID
- Name, description and business tags
- Lineage information (e.g. source system)
- Data schema (fields, types etc...)
By integrating with your version control system, trac ensures that a complete and immutable version history is maintained for every model. The automated governance report for model objects includes:
- Object & version ID
- Name, description and business tags
- Lineage data (e.g. source repo, commit hash)
- Model schema (inputs, outputs, parameters)
Each Flow is the blueprint of a complex calculation which trac can run on demand — they exist only as immutable TRAC metadata. The governance report for a Flow includes:
- Object & version ID
- Name, description and business tags
- Lineage data (e.g. creation date)
- Visual representation of the graph
The instruction used to orchestrate a calculation also acts as its comprehensive audit record. A Job governance report includes:
- Job ID
- Lineage data (run on, run by)
- Details of the Flow that was used
- Schemas and object IDs for all the models and data inputs
- Numerical values of all the runtime parameters
- Schemas and object IDs for all outputs the job generated
Repeatable
Faithfully recreate any prior calculation in minutes
A complete metadata representation of all the platform objects, processes, and calculations ensures you have comprehensive, governance-ready documentation at the your fingertips, with no manual input needed.