Free & open source
FinOps Tools
The cloud bill arrives by account and service, not by team: finops-skills guides your
coding agent to build cost by team, in your own stack. AI assistants bill by the token, but
the value lands in commits: claude-budget and Copilot Budget put a cost trailer on every
AI-assisted commit, so you can answer "what did this feature cost?" straight from
git log.
-
finops-skills
Agent skill · cloud cost by team
Guides your coding agent to build cloud cost attribution by team on FOCUS exports, in your own stack.
It scans your bill before asking anything, interviews you only for what the data can't show, then builds phase by phase from specs and golden test cases, with a verification gate after each phase. Try it on a sample bill first, with no credentials.
- Claude Code or any SKILL.md agent
- AWS, Azure, Google Cloud
- Markdown and YAML, no code
- MIT-licensed
-
claude-budget
CLI · Claude Code cost per commit
Puts a price on every commit made with Claude Code, with no editor in the loop.
A single static Go binary reads the session transcripts Claude Code already writes, prices the tokens against an embedded Anthropic rate card, and attaches a Claude-Cost trailer to each commit. One setup command installs the hooks; then you commit as usual.
- Single static Go binary
- No daemon, no account linking, no telemetry
- MIT-licensed
-
Copilot Budget
VS Code extension · Copilot cost per commit
Puts a price on every commit made with GitHub Copilot.
It reads per-span token counts from Copilot Chat's local telemetry database and appends an AI-cost trailer to your commit message via a prepare-commit-msg hook — so the receipt lives in git history, next to the code it paid for.
- VS Code extension
- Works on any Copilot plan
- MIT-licensed
- No accounts, no telemetry sent anywhere
The build finops-skills walks your agent through, from the raw bill to ownership rules in Git, is in Map Your Cloud Bill to Your Teams: Build Your Own FinOps Tool.
The methodology behind claude-budget and Copilot Budget — why the commit is the right place to attribute AI cost, and how to read the numbers once you have them — is in FinOps for AI: Track What Your Code Actually Costs Per Commit.
Rolling out AI cost tracking across a whole engineering org? That's a conversation.