# Getting started

> Sign in, use the data in the app, make an API key, and connect an AI agent.

## Create an account

[Sign up](https://getcadastre.com/signup) with your name and email address. Cadastre emails you a link; the page it opens is where you choose a password and accept the Terms of Service and Privacy Policy. After that, [sign in](https://getcadastre.com/login) with your email and password, or with a passkey once you have added one in Settings.

## Use the data in the app

1. Create a project for the client or question you are working on.
2. On **MSA Selection**, narrow the geographic scope, set requirements (for example a minimum population or a maximum unemployment rate) and add criteria in plain language. Each criterion is matched to the measures described in [Datasets](https://getcadastre.com/docs/datasets), and you approve it before it counts.
3. The register ranks the metros that pass, and the map shades them by score. Save a snapshot to keep a ranking exactly as it was computed.
4. On **Reports**, download the snapshot as a PDF or a PowerPoint deck, or its data as CSV or Excel. Each lists its data sources with the credit they require.

## Programmatic access

The read-only [API](https://getcadastre.com/docs/api), the [MCP server](https://getcadastre.com/docs/mcp) and the [CLI](https://getcadastre.com/docs/cli) are live in beta, free while they are, and need no account for the public datasets: call them with no credential, at 30 requests a minute and 2,000 a day per network address. (The app itself needs an account.)

```sh
curl "https://getcadastre.com/api/v1/metros?q=austin"
```

For higher limits, sign in, with an account that has API access turned on. A new account does not have it yet: write to support@getcadastre.com to ask for it (API access may become part of a paid plan later). Until it is on, Settings makes no API keys for the account and apps cannot be connected to it. There are two ways to sign in:

### API keys

For your own scripts and servers. In [Settings](https://getcadastre.com/account), under **API keys**, make a key (you are asked for your password). It starts `cad_key_`, is shown once, and acts as you in the account you were working in. Send it as a bearer token:

```sh
curl -H "Authorization: Bearer $CADASTRE_API_KEY" https://getcadastre.com/api/v1/me
```

Keep a key secret, as you would a password; revoke it in the same place. Keys are for scripts you run, not for software or pages other people can inspect.

### OAuth

For AI agents and other apps that sign in on your behalf. The app opens your browser, you sign in to Cadastre, choose the account and approve; the app keeps its own tokens and refreshes them, so there is no key to copy. Discovery starts at [https://getcadastre.com/.well-known/oauth-authorization-server](https://getcadastre.com/.well-known/oauth-authorization-server). Disconnect an app in Settings under **Authorized apps**. To set up a particular agent (Claude Code, Codex, Cursor, GitHub Copilot in VS Code, Claude, ChatGPT, Gemini CLI, or OpenCode), see [Connect an AI agent](https://getcadastre.com/agents) and the [MCP server](https://getcadastre.com/docs/mcp) page.

Each key and each approved app has its own rate limits, and every answer carries the credit its sources require: see [API](https://getcadastre.com/docs/api).

### The command line

The [Cadastre CLI](https://getcadastre.com/docs/cli), `cadastre`, runs screens, rankings and comparisons in the terminal, with the attribution printed under every answer. It works without signing in; `cadastre login` signs in with OAuth in the browser (or takes an API key) for higher limits:

```sh
curl -fsSL https://raw.githubusercontent.com/waratuman/cadastre-cli/main/install.sh | sh
cadastre metros --state TX
cadastre login   # optional: higher limits
```

## For AI agents and tools

These docs are written to be read by agents as well as people:

- Every page is Markdown at its own address plus `.md`, for example [https://getcadastre.com/docs/datasets.md](https://getcadastre.com/docs/datasets.md).
- [/llms.txt](https://getcadastre.com/llms.txt) lists every page with a one-line description, per the [llms.txt convention](https://llmstxt.org/).
- [/llms-full.txt](https://getcadastre.com/llms-full.txt) is the whole of the docs as one Markdown file.

```sh
curl https://getcadastre.com/llms.txt
curl https://getcadastre.com/docs/datasets/acs.md
```

Pages describe the data and its rights; they carry no data values.