Getting started

Install spendlens and its two engines, point it at an AWS account, and run your first checks. Allow about ten minutes.

Requirements

1. Install

# engines
brew install turbot/tap/steampipe turbot/tap/powerpipe      # macOS / Linuxbrew
# or: sudo /bin/sh -c "$(curl -fsSL https://steampipe.io/install/steampipe.sh)"
#     sudo /bin/sh -c "$(curl -fsSL https://powerpipe.io/install/powerpipe.sh)"

steampipe plugin install aws

# spendlens (the [ai] extra adds the AI assistant's SDKs)
python3 -m venv .venv && . .venv/bin/activate
pip install "spendlens[ai] @ git+https://github.com/devops-monk/spendlens.git"

Steampipe refuses to run as root. On servers and in CI, run spendlens as a normal user.

2. Try it without AWS

Every command can replay bundled sample data, so you can explore before connecting an account:

spendlens tags --source demo --summary
spendlens waste --source demo
spendlens cost diff --source demo
spendlens report --source demo --out demo-report      # opens as demo-report/report.html

3. Connect your AWS account

spendlens uses the AWS credential chain, so whatever works for the AWS CLI works here:

aws sso login --profile finops            # or: export AWS_PROFILE=finops, or an assumed-role session
steampipe service start

Use --profile on any command, or set aws.profile in finops.yaml. For several accounts, configure a Steampipe aggregator connection and set steampipe.connection to its name.

4. Describe your standards

spendlens init          # writes finops.yaml

Edit the mandatory tags, allowed instance families, budgets and AI provider. See Configuration for every option.

5. Check your setup

spendlens doctor

doctor checks Python, Steampipe, Powerpipe, the AWS plugin, your AWS identity, the Steampipe service, Cost Explorer access and the AI assistant setup. Each failed check prints its fix. spendlens doctor --offline runs only the local checks and makes no network calls.

6. Run it

spendlens tags --summary
spendlens rightsize --min-saving 20
spendlens waste
spendlens cost diff
spendlens report --out reports        # everything at once, plus reports/report.html

Every live run is saved to your local history (~/.local/share/spendlens/spendlens.db), so you can re-render any report offline with --source last.

Next

  • Configuration: tune thresholds to your company's standards
  • AI assistant: ask questions about your spend
  • CI/CD: run on a schedule and post summaries to Slack or Teams