JMESPath is a query language for JSON, used most widely by the AWS CLI and boto3 in the --query option. An expression transforms a JSON document into a smaller, reshaped result, letting you extract exactly the fields you need without piping through jq.

How is JMESPath different from JSONPath?

Both query JSON, but JMESPath has a stricter, fully specified grammar with first-class projections, pipe expressions, multiselect lists and hashes, and a standard function library. JSONPath dialects vary between implementations, whereas JMESPath behaves identically everywhere the spec is implemented.

Which features does this tester support?

It supports field access, indices and slices, list and object projections [] [*] *, filter projections [?expr], pipes |, comparison and logical operators, literals in backticks, multiselect lists and hashes, and common functions including length, keys, values, sort, sort_by, min, max, sum, avg, join, contains, starts_with, to_string and to_number.

Does this tool match the AWS CLI exactly?

For the common subset most --query filters use, yes — the same expression produces the same shape. A handful of rarely used spec corners are simplified, so always confirm complex expressions against the real CLI before relying on them in automation.

Is my JSON uploaded anywhere?

No. The JSON is parsed and the expression evaluated entirely in your browser with JavaScript. Nothing is sent to a server, so you can safely test output that contains account IDs or other sensitive values.

JMESPath Tester — Gera Tools

JMESPath is the JSON query language built into the AWS CLI, boto3 and many other tools. When you pass --query to an aws command, the string you supply is a JMESPath expression that filters and reshapes the JSON response. This tester evaluates those expressions live against your own JSON so you can perfect a --query before running it against a real account.

How it works

A JMESPath expression is read left to right. The most important building block is the projection, which applies the rest of the expression to every element of a list or object:

Reservations[].Instances[].InstanceId

Here Reservations[] flattens the list of reservations, Instances[] flattens the instances inside each, and .InstanceId is then applied to every instance, producing a flat list of IDs. A filter projection narrows a list:

Instances[?State == `running`].InstanceId

keeps only the instances whose State equals the literal running (literals go inside backticks), then projects out their IDs. A pipe stops a projection and starts a new expression on its result:

Instances[].Cpu | max(@)

projects every CPU value, then max(@) runs once on the resulting array, where @ means “the current value”.

Functions and multiselect

The standard function library lets you reshape and aggregate: length(@), sort_by(Instances, &Cpu), join(', ', Names), contains(Tags, 'prod'), sum(Costs) and more. A multiselect hash rebuilds objects with only the fields you want:

Instances[].{id: InstanceId, cpu: Cpu}

returns a list of small objects, each containing just id and cpu.

Tips

Wrap literal values in backticks: State == \running`, Count == `3`. Bare quotes ’text’` are raw strings and also work in comparisons.
Use [] to flatten nested lists and [*] to project a single level. They behave differently when the data is nested, so reach for [] when AWS wraps results in an extra list (as it does with Reservations).
Build the expression incrementally. Start with Reservations[], confirm the shape, then append the next segment.