kq

KQL CLI — query Azure Data Explorer (Kusto) from the command line.

Like jq for JSON, but for Kusto/KQL. Run raw KQL, keep a git-versioned library of parameterized queries, and pipe results straight into your shell.

Installation

pip install kql-cli

The command you run is kq. The PyPI package is named kql-cli because kq was already taken on PyPI by an unrelated project.

Or from source:

git clone https://github.com/cptfinch/kq.git
cd kq
pip install -e .

Quick Start

# Configure your cluster
kq config set default_cluster https://mycluster.westeurope.kusto.windows.net
kq config set default_database mydb

# Authenticate
kq auth login

# Run queries
kq "MyTable | take 5"                    # Raw KQL
kq list                                   # List saved queries
kq run examples.sample MyTable 10         # Run saved query

Configuration

Config is stored in ~/.config/kq/config.yaml:

default_cluster: https://mycluster.westeurope.kusto.windows.net
default_database: mydb

clusters:
  prod:
    url: https://prod.westeurope.kusto.windows.net
    database: proddb
  dev:
    url: https://dev.westeurope.kusto.windows.net
    database: devdb

Configure via CLI:

kq config show                              # Show current config
kq config set default_cluster <url>         # Set default cluster
kq config set default_database <db>         # Set default database
kq config add-cluster prod <url> --database proddb  # Add named cluster

Commands

Saved Queries

Queries are loaded from (in priority order):

1. ./.kq/ - Project-local queries
2. ~/.config/kq/queries/ - User queries
3. Bundled examples

Query Format

Create YAML files in ~/.config/kq/queries/:

# ~/.config/kq/queries/myqueries.yaml
name: myqueries
description: My custom queries

queries:
  - name: recent
    description: Get recent records
    safety: safe
    parameters:
      - name: table
        description: Table name
        required: true
      - name: hours
        description: Hours to look back
        default: "24"
    query: |
      {table}
      | where Timestamp > ago({hours}h)
      | order by Timestamp desc
      | take 100
    example: "MyTable 24"

Then run:

kq list                        # Shows myqueries.recent
kq show myqueries.recent       # Show details
kq run myqueries.recent Events # Run with parameters

Output Formats

kq "MyTable | take 5" -f table    # Default - human readable
kq "MyTable | take 5" -f json     # JSON array
kq "MyTable | take 5" -f csv      # CSV

Authentication

Supports (in priority order):

1. Service Principal - Set AZURE_CLIENT_ID, AZURE_CLIENT_SECRET, AZURE_TENANT_ID
2. Azure CLI - Run az login first
3. Device Code - Interactive browser login (tokens cached ~90 days)

Query Safety

Queries have a safety level:

• safe - Queries with proper time/scope filtering
• caution - May scan significant data, use carefully
• dangerous - Can scan entire tables, requires explicit filtering

Always filter by time first:

// Good - filters first, cheap
MyTable | where Timestamp > ago(1d) | where Category == 'Error'

// Bad - scans everything, expensive
MyTable | where Category == 'Error'

Why kq?

• LLM-native - Works seamlessly with Claude Code, Copilot, etc.
• Portable - Same queries work across clusters
• Versionable - Git-controlled query libraries
• Unix-friendly - Pipes, scripts, automation
• Personal queries - User queries never overwritten by updates

Development

git clone https://github.com/cptfinch/kq.git
cd kq
python -m venv .venv && . .venv/bin/activate
pip install -e ".[dev]"

pytest            # run tests
ruff check .      # lint
python -m build   # build sdist + wheel

CI runs lint + tests across Python 3.9–3.13 on every push and pull request.

Releasing

Releases publish to PyPI automatically via Trusted Publishing (OIDC — no tokens stored in the repo). To cut a release:

1. Bump __version__ in src/kq/__init__.py and update CHANGELOG.md.
2. Tag and push: git tag v1.2.3 && git push origin v1.2.3.

The release.yml workflow builds the artifacts and publishes them. This requires a one-time PyPI setup: configure kql-cli's trusted publisher to point at this repository, workflow release.yml, environment pypi.

License

MIT — see LICENSE.