Research data,
checkpointed.
TurtleTech Data keeps datasets, models, and simulation artifacts attached to reproducible checkpoints. The Tardigrade native client and protocol are the default access path; DVC compatibility supports established projects.
$ tt auth login $ tt init $ tt add data/training-set/ $ tt save "baseline dataset" Tardigrade saved and pushed your changes.
Tardigrade native client
The tt workflow uses the native TurtleTech Data protocol to record datasets and models with their checkpoints.
DVC compatibility
Existing DVC projects can use the managed compatibility endpoint while moving toward the native workflow.
App-scoped capacity
The TurtleTech Data allowance is separate from Ridley and every other TurtleTech app.
EU-hosted storage
Research objects are stored in an EU object-storage region and accessed over encrypted HTTPS.
Why version your data?
You ran an analysis and a collaborator needs to reproduce it. Git tracks source well, but large datasets need their own storage and history. TurtleTech Data connects those objects to named checkpoints through the Tardigrade protocol without making DVC the product's underlying model.
The Turing Way has a good overview of why data version control matters and how data versioning fits into a reproducible research workflow.
Preview plan catalogue
These plans are available only to manually provisioned preview accounts. Creating a TurtleTech ID does not activate TurtleTech Data storage.