CoreWeave AI Object Storage now supports the RenameObject API, providing atomic, server-side rename operations for objects within the same bucket. RenameObject updates index entries instead of copying data, so rename operations complete in milliseconds regardless of object size.Documentation Index
Fetch the complete documentation index at: https://docs.coreweave.com/llms.txt
Use this file to discover all available pages before exploring further.
Overview
The RenameObject API provides a simpler and more efficient alternative to copy-and-delete workflows. Key benefits include:- Atomic operation: No window where both keys exist simultaneously
- No data movement: Updates metadata only, avoiding copy latency and temporary storage duplication
- Metadata preservation: Encryption settings, storage billing tier, checksums, and other attributes are preserved
- Precondition support: ETag and time-based conditions guard against race conditions and accidental overwrites
Use cases
| Use case | Description |
|---|---|
| Checkpoint promotion | Rename temporary checkpoints to final names after training runs complete |
| Batch finalization | Move artifacts from staging prefixes to final locations after validation |
| Log rotation | Atomically “lock” time-based prefixes by renaming to finalized paths |
Limitations
- Same bucket only: cross-bucket rename is not supported
- Versioning not supported: buckets with versioning enabled or previously enabled cannot use RenameObject
- Single object per request: prefix-level rename must be implemented as client-side batches
Learn more
- See the Rename Objects guide for an end-to-end walkthrough with examples
- See the RenameObject API reference for S3 compatibility details
- See conditional writes (March 23, 2026) for the same precondition pattern on PutObject, CompleteMultipartUpload, and CopyObject