Docs
Caching
Contigra’s cache, the Velocity Engine, avoids repeating work that has not changed.
How it works
Cache entries are keyed by content, so a result is reused only while its inputs are unchanged and is invalidated automatically when files change. Results persist across runs.
Options
| Option | Description |
|---|---|
| Eviction policy | How entries are removed when the cache is full: LRU, LFU, or usage-based. |
| Maximum size and age | Upper bounds on total size and entry lifetime. |
| Compression | Compress stored entries to save space. |
| Project isolation | Give each project its own cache boundary and size limit. |
Managing the cache
contigra cache statscontigra cache --help lists the other cache commands. Disable caching for a single run when you need a clean build.