Hydra currently uses Perforce as its production version-control system. The actual integration is concentrated in hydraperforcewatcher: it polls submitted changelists, identifies files changed under a build path, exports those files, uploads the build, and notifies downstream services.
Epic has released Lore, an MIT-licensed, binary-first version control system designed for large game and entertainment projects:
https://dev.epicgames.com/documentation/en-us/unreal-engine/lore-version-control-in-unreal-engine
Lore may fit Hydra well because it offers content-addressed storage, fragment-level deduplication, sparse workspaces, cheap branches, self-hosting, and a Go SDK.
Do not replace Perforce yet. Add Lore only as an experimental backend and retain Perforce as the production default.
As of September 2026, Lore is pre-1.0 and its APIs and protocols may still change. File locking is advisory rather than enforced, and Epic lists the native Unreal Editor plugin on its 2027 roadmap. These are meaningful risks for binary-heavy artist workflows.
The watcher currently depends on server connectivity, revision discovery and path filtering, listing changed files, exporting files at a revision, and enumerating streams and depots. Lore appears to provide equivalents through its CLI/API. The main unknown is efficient, structured discovery of revisions that affect a configured build path.
If successful, introduce a neutral repository interface above the existing Perforce client, with Perforce and Lore implementations. Replace Perforce-specific state and payload concepts with backend-neutral revision identifiers while retaining compatibility for downstream changelist consumers.
A pilot is approximately 6–10 engineer-weeks across the watcher adapter, neutral state/schema changes, dashboard compatibility, server deployment/provisioning, and operational testing. A full production replacement requires additional migration tooling and hardening.