RepliLedger

RepliLedger is a web app (with a lightweight CLI) that continuously verifies whether a research repo can be rebuilt and rerun from scratch. Connect a GitHub/GitLab repository, and it spins up clean, ephemeral environments to execute the project’s pipeline (e.g., Snakemake/Nextflow/Make), captures exact dependency locks, seeds, hardware info, and produces a signed “repro report” with pass/fail status, runtime, and artifact hashes. It also generates a minimal “repro bundle” (container + lockfiles + run manifest) that can be cited in papers and shared with reviewers. This is not magic: it won’t fix bad science or missing data, but it will surface the real blockers—undocumented steps, nondeterminism, hidden manual edits—before submission. Teams get a dashboard that tracks reproducibility drift over time, so a project doesn’t rot between revisions.