S3-compatible storage with built-in search.S3-compatible storage with built-in search.
Upload and download with S3 clients, then use Keldra's native API to find objects by indexed metadata, JSON fields, text or vectors. Your original files stay in the same store.
Use your existing workflows
Upload and download through supported S3 requests and SDKs.
Query your data
Find objects by declared metadata, JSON fields, words or embeddings.
Keep fewer moving parts
Object storage and the indexes your application needs share one platform.
Store as usual. Index what matters.
Keep S3 file transfers. Add native queries.
Point compatible tools, scripts and applications at your Keldra endpoint. Use path-style addressing and SigV4 signing; your application's client ID and secret are its access key and secret key.
Define the indexes you need before uploading or importing objects. Supported incoming content becomes searchable through Keldra's native query API; your S3 download workflow still returns the original file.
From S3 uploads to indexed queries
Your applications, SDKs and CLI tools
Upload and download objects
Your existing workflows
Keldra
S3-compatible storage with declared indexes
Search and query
Explore metadata
Work with your data
Your applications, SDKs and CLI tools
Upload and download objects
Your existing workflows
Keldra
S3-compatible storage with declared indexes
Search and query
Explore metadata
Work with your data
Useful search without another pipeline
Adding search to an object store often leaves the application to move and reconcile data between services. Keldra brings supported indexing closer to the objects it describes.
No separate crawler just to index supported incoming objects
Fewer indexing services and queues to operate
Object changes feed the indexes in the storage platform
Keep the original files, not a replacement export format
Use the S3 operations you need
Common file operations work through the S3 surface. Rich field queries, aggregations and native streaming use Keldra's own API.
Use S3-compatible clients for everyday file operations and the native Keldra client for streaming large files, indexed queries and aggregations.
Configure your S3 client with the Keldra endpoint, path-style bucket addressing and SigV4 credentials. The operation table below shows the requests to use.
Supported S3 operations
Create buckets, upload and download files, retrieve metadata and page through objects with these S3 operations.
Operation
Status in Keldra
Notes
CreateBucket · PUT bucket
Supported
Create a bucket using path-style requests and SigV4.
HeadBucket · HEAD bucket
Supported
Check that a bucket exists.
ListObjectsV2
Supported
Page through objects in a bucket.
PutObject · PUT
Supported
Upload an object.
HeadObject · HEAD
Supported
Retrieve object metadata without its bytes.
GetObject · GET
Supported
Download the object, including supported single byte ranges.
DeleteObject · DELETE
Supported
Delete an object.
Store files through S3. Search their indexed data.
Keep S3 tools and add useful queries with Keldra.
Your data stays in your environment.
Keep the tools that use supported S3 operations.
Spend less time joining storage and search services.