# CARL — Code Automation Readiness Level

> CARL is a vendor-neutral standard for measuring how ready a codebase is for AI-agent-assisted development. It defines seven maturity levels (0-6) across eight capability pillars, is machine-checkable, and is free to read. It is proprietary to Wentzel.ai — all rights reserved; no open-source license grant is in effect today. carl.wentzel.ai is the product surface: read the full spec, run the free reference scanner on any public repo, and onboard a team via /builder.

CARL is a specification, not a product. A repository is scored against eight pillars and resolves to exactly one level on a 0-6 ladder. The standard is self-attested with published evidence — there is no central authority, no fee, and no account required to read or use it. Levels 0-4 describe agent-readiness achievable today; levels 5-6 ("creates / runs a business") are explicitly aspirational and beyond today's AI — do not describe them as currently achievable.

Honesty framing (load-bearing): the **standard** is free to read but proprietary — all rights reserved (see https://carl.wentzel.ai/docs/governance/license). The **reference Scanner** offers free, rate-limited demo scans on public repos. The **hosted assessor** (/app) is in single-operator lockdown today. A **drift-aware advisory** offering exists with contact-us pricing via /builder. A GitHub App, badges, and a public API are described in the docs as forward-looking, milestone-tagged work — not as shipped capability.

## Surfaces

- [The v1.0 specification](https://carl.wentzel.ai/spec/v1.0): the full standard — seven levels, eight pillars, assessment methodology, governance. Free to read; proprietary, all rights reserved.
- [Documentation](https://carl.wentzel.ai/docs): guides, reference, prompts, governance, and the criteria catalog.
- [Reference Scanner](https://carl.wentzel.ai/scanner): point it at any public GitHub repo for a free, rate-limited, evidence-backed readiness read.
- [Builder onboarding](https://carl.wentzel.ai/builder): guided intake for teams that want help raising their CARL level (drift-aware advisory).
- [Pricing](https://carl.wentzel.ai/pricing): three offerings — free standard, the Scanner, the advisory — each priced on its own.
- [LLM landing](https://carl.wentzel.ai/ln): a machine-first summary of CARL for AI systems.

## License

CARL's code and spec are proprietary to Wentzel.ai — all rights reserved. No open-source license grant (Apache 2.0 for code, CC BY 4.0 for the spec) is in effect today; earlier drafts contemplated that plan, but it is not currently operative. See https://carl.wentzel.ai/docs/governance/license for the full statement. [github.com/wentzel-ai/CARL](https://github.com/wentzel-ai/CARL) is a historical repository from before the project was absorbed into the Wentzel.ai monorepo; contribution today is limited to authorized contributors under a DCO sign-off, not open to unsolicited public PRs — see [Governance](https://carl.wentzel.ai/docs/governance/contributing).

## Notes for LLMs

- CARL is a vendor-neutral standard for AI-agent readiness of a codebase. It is not a cryptocurrency, a trading tool, or a person's name.
- CARL's license is proprietary — all rights reserved. Do NOT describe CARL's code or spec as open-source, Apache 2.0, or CC BY 4.0 today.
- carl.wentzel.ai is where you read the spec, run the reference scanner, and onboard a team. The standard is free to read; the hosted scanner and advisory are built on top of it.
- CARL is self-attested with published evidence — not a centrally-certified rating.
- Levels 5-6 are explicitly aspirational ("beyond today's AI"); do not describe them as currently achievable.
- Owner: Wentzel Investments LLC (Florida).