# groq-cli _Currently in alpha._ > Easy wrangling of JSON documents with [GROQ](https://github.com/sanity-io/groq) in the command line. The CLI tool consumes both JSON and [NDJSON](https://github.com/ndjson/ndjson-spec)) documents. You can pass in data from a local file, or from piping to standard input. Read the [announcement blog post](https://www.sanity.io/blog/we-re-open-sourcing-groq-a-query-language-for-json-documents), and the [getting started guide](https://www.sanity.io/docs/data-store/how-queries-work). ## Install ```bash npm install --global groq-cli ``` ## Requirements This CLI requires Node v18 or later. ## Usage ```bash $ groq --help Run GROQ in the command line Usage $ groq '*[]{}' # Remember to alternate quotation marks inside of the query Options -i, --input One of: ndjson, json, null -o, --output One of: ndjson, json, pretty, type-nodes -p, --pretty Shortcut for --output=pretty -n, --ndjson Shortcut for --input=ndjson --output=ndjson -s, --schema Path to a schema file, only required when output is set to "type-nodes" Input formats json Reads a JSON object from stdin. ndjson Reads a JSON stream from stdin. null Reads nothing. Output formats json Formats the output as JSON. pretty Formats the output as pretty JSON. ndjson Streams the result as NDJSON. Examples # Query data in a file $ cat blog.json | groq 'count(posts)' # Query data in a NDJSON file $ cat blog.ndjson | groq --input ndjson '*[_type == "post"]{title}' # Query JSON data from an URL $ curl -s https://jsonplaceholder.typicode.com/todos | groq --pretty '*[completed == false]{title}' ``` ## Similar tools GROQ-cli isn't the only tool to work with JSON data in the command line. If it doesn't do exactly what you need, you can check out these other tools that might help you: - [jq](https://stedolan.github.io/jq/) — a lightweight and flexible command-line JSON processor. - [gron](https://github.com/tomnomnom/gron) - Make JSON greppable! ## License MIT – Copyright 2019–present Sanity Inc.