Store Checkpoint Data - Entire

Documentation Index

Fetch the complete documentation index at: /llms.txt

Use this file to discover all available pages before exploring further.

Entire stores checkpoint data separately from your normal source branches. By default, that data is written to a dedicated Git branch in your code repository named entire/checkpoints/v1. Your code stays on your source branches, while prompts, transcripts, session context, and checkpoint metadata live in checkpoint storage. Checkpoint storage involves two choices:

• How checkpoint data is organized inside Git: on one shared branch or in one Git ref per checkpoint.
• Where checkpoint data is synchronized: with your code, in a separate repository, or nowhere outside your machine.

Choose where checkpoints go

By default, checkpoints stay with the repository they describe. This keeps the code and the context behind the code together. The checkpoint backend controls how checkpoint data is organized inside Git. The options below separately control which repository receives that data and whether it is pushed. You can change that default when checkpoint history should follow a different storage or access boundary:

• Keep checkpoints with the code repository
• Push checkpoints to a separate repository
• Keep checkpoints local on your machine

Where storage settings are saved

Storage settings are written to .entire/settings.json by default, so they can be shared with the repository. For personal preferences that should not affect the rest of the team, use the --local flag to write to .entire/settings.local.json instead. Local settings apply only to your clone and take precedence over project settings. See Local, Project, and Global Settings for details.