pick

pick-cli.pages.dev

Extract values from anything — JSON, YAML, TOML, .env, HTTP headers, logfmt, CSV, and more.

cargo install pick-cli

pick auto-detects the input format and lets you extract, filter, and transform values using a unified selector syntax. Think of it as jq for all config formats — no more juggling jq, yq, grep, awk, and cut.

Quick Start

# JSON
curl -s https://api.github.com/users/octocat | pick login
# octocat

# .env
cat .env | pick DATABASE_URL
# postgres://localhost:5432/mydb

# YAML
cat config.yaml | pick server.port
# 8080

# TOML
cat Cargo.toml | pick package.version
# 0.1.0

# HTTP headers
curl -sI https://example.com | pick content-type
# text/html; charset=UTF-8

# logfmt
echo 'level=info msg="request handled" status=200' | pick status
# 200

# CSV
cat data.csv | pick '[0].name'
# Alice

Selectors

Pipes & Filters

Chain operations with the pipe operator (|), filter with select(), and transform with builtins:

# Filter: find expensive items
cat data.json | pick 'items[*] | select(.price > 100) | name'

# Regex: match patterns
cat data.json | pick 'items[*] | select(.email ~ "@gmail\\.com$") | name'

# Boolean logic: and, or, not
cat data.json | pick 'users[*] | select(.age >= 18 and .active == true) | name'

# Builtins: keys, values, length
cat config.json | pick 'keys()'
cat config.json | pick 'dependencies | length()'
cat data.json | pick 'items[*].name | length()'

# Chain multiple stages
cat data.json | pick 'items[*] | select(.price > 50) | name | length()'

Filter Operators

Builtins

Mutation

Modify data in-place with set() and del():

# Set a value
cat config.json | pick 'set(.version, "2.0")' --json

# Delete a key
cat config.json | pick 'del(.temp)' --json

# Chain mutations
cat data.json | pick 'set(.status, "active") | del(.temp)' --json

# Mutate then extract
cat config.json | pick 'set(.version, "2.0") | version'

Output Formats

Convert between formats with --output:

# JSON to YAML
cat data.json | pick -o yaml

# JSON to TOML
cat data.json | pick 'config' -o toml

# Always output JSON
cat data.json | pick 'name' --json

Streaming (JSONL)

Process newline-delimited JSON (JSONL) line-by-line with --stream:

# Extract from each line
cat events.jsonl | pick 'user.name' --stream

# Filter streamed data
cat logs.jsonl | pick 'select(.level == "error") | message' --stream

# Stream from file
pick -f events.jsonl 'timestamp' --stream

Flags

Examples

Pipe-friendly

# Get the current git user's repos
curl -s https://api.github.com/users/octocat/repos | pick '[*].name' --lines

# Check if a key exists before using it
if cat config.json | pick database.host --exists; then
  DB_HOST=$(cat config.json | pick database.host)
fi

# Extract with a fallback
cat config.yaml | pick server.port --default 3000

# Count results
echo '[1,2,3,4,5]' | pick '[*]' --count
# 5

Real-world

# Docker container status
docker inspect mycontainer | pick '[0].State.Status'

# Kubernetes pod IPs
kubectl get pods -o yaml | pick 'items[*].status.podIP' --lines

# Find all IDs recursively in any JSON
cat response.json | pick '..id'

# Cargo.toml dependencies
pick -f Cargo.toml dependencies.serde.version

# GitHub API: filter Rust repos with 100+ stars
curl -s https://api.github.com/users/octocat/repos | \
  pick '[*] | select(.language == "Rust" and .stargazers_count > 100) | name'

# Terraform outputs
terraform output -json | pick '..value'

# npm: count dependencies
cat package.json | pick 'dependencies | keys() | length()'

# .env database URL for a script
export DB=$(cat .env | pick DATABASE_URL)

Supported Formats

Auto-detection works in most cases. Use -i to override when the input is ambiguous.

Install

Cargo (Rust)

cargo install pick-cli

Homebrew (macOS/Linux)

brew install aryanbhosale/pick/pick

npm

npm install -g @aryanbhosale/pick

Snap (Linux)

snap install pick-cli

Chocolatey (Windows)

choco install pick

WinGet (Windows)

winget install aryanbhosale.pick

Docker

echo '{"foo":"bar"}' | docker run -i ghcr.io/aryanbhosale/pick foo

GitHub Releases

Download pre-built binaries from Releases — macOS (ARM/x64), Linux (x64/ARM64), and Windows (x64).

From source

Requires Rust 1.85+:

git clone https://github.com/aryanbhosale/pick.git
cd pick
cargo install --path .

Contributing

Contributions are welcome! Here's how to get started:

1. Fork the repository
2. Create a feature branch: git checkout -b my-feature
3. Make your changes
4. Run the tests: cargo test
5. Commit and push: git push origin my-feature
6. Open a pull request

Development

# Run all tests (866 tests)
cargo test

# Run a specific test
cargo test test_name

# Build release binary
cargo build --release

# The binary will be at target/release/pick

Project Structure

src/
  main.rs          Entry point, stdin/file reading, streaming
  lib.rs           Orchestration and format routing
  cli.rs           CLI argument definitions
  error.rs         Error types
  selector/        Selector engine (modular)
    types.rs       AST types (Expression, Pipeline, Selector, Filter, etc.)
    parser.rs      Hand-rolled recursive descent parser
    extract.rs     Path traversal and pipeline execution
    filter.rs      Filter evaluation (select, comparisons, regex)
    manipulate.rs  set() and del() operations
  detector.rs      Format auto-detection heuristics
  output.rs        Output formatting (plain, JSON, YAML, TOML)
  streaming.rs     JSONL streaming processor
  formats/         Per-format parsers
    json.rs, yaml.rs, toml_format.rs, env.rs,
    headers.rs, logfmt.rs, csv_format.rs, text.rs
tests/
  integration.rs   CLI integration tests

Issues

Found a bug or have a feature request? Open an issue.

License

MIT