Filter JSON fields. Group and total the matches.
Filter JSON fields. Group and total the matches.

Select orders by status, date or amount and return counts and sums alongside matching records. Declared field indexes keep those calculations inside Keldra rather than in application-side document loops.
For teams that run orders, tasks and inventory, typed JSON indexing turns declared fields into useful answers. Query exact values, dates, amounts or status, calculate totals, and retrieve the records you need.
Find
Locate the right records across your JSON.
Understand
Get counts and totals for real decisions.
Keep control
Your data, your permissions, on your infrastructure.
Structured answers from real JSON.
A COMMON SCENARIO

Query orders, due tasks and stock records.

Your business data already lives in JSON — orders, tasks, inventory and more. Index the fields that help you get the answers you need, when you need them.

Orders

Find recent orders, filter by amount or status, check totals and see what is pending.

Tasks

Find ready, due tasks, then use keyed filters to compare each tenant's running-container count with its capacity setting.

Inventory

Select stock records by their declared item fields and calculate totals without fetching every payload.

Find recent unshipped orders and total their value.

A stored order stays a JSON document. Its indexed fields let your application ask useful questions.

JSON document: order

{ "order_id": "ORD-1042", "customer": {"name": "Acme Tools", "tier": "business"}, "items": [ {"sku": "TL-001", "name": "Drill", "qty": 2, "price": 129.00}, {"sku": "TB-004", "name": "Bits set", "qty": 1, "price": 24.50} ], "total": 282.50, "status": "processing", "placed_at": "2026-10-12" }

The question you want answered

Show orders from the last 30 days with a total over £100 that are not yet shipped, newest first. Return their count and the sum of their totals.
Provision the date order and the exact, range and grouping capabilities the query uses. Filters, ordering and aggregations operate over the JSON fields you indexed.

Index for the questions you ask.

Choose which fields to index based on the questions your team needs to answer.
order_id
customer.tier
status
total
placed_at
items.sku
Exact lookup, range, grouping and physical ordering are distinct capabilities. Storing a field alone does not make every operation on it indexed.
A range on a multivalued field matches when an indexed value satisfies that range. It does not turn an entire array into one number.
For efficient deep paging, provision the complete physical order. A field being searchable does not make every arbitrary sort cheap.

Query within what the user can see.

Supply the custom realm, subject and relation for each query input. Keldra applies visibility before producing hits, groups and aggregations.

Authorised objects returned

Tasks, running containers and settings can use different relations within the same keyed query.

Filter, sort and aggregate

Use declared indexed fields for boolean conditions, ranges, counts, sums and other supported aggregates.
Generic object writes use application and bucket permissions; they do not automatically infer a custom end user's edit relation.

Query the JSON records your application already stores.

Useful answers from the records your application already stores.