@aglyn/cli

Read and script any site built on Aglyn — pages as Markdown, llms.txt, OpenAPI, and the REST API.

npm install -g @aglyn/cli

Most of it needs no account

Every Aglyn site publishes a public, anonymous surface. These commands work against any Aglyn site — including one you do not own — with no key, no plan and no sign-in:

aglyn read https://example.com/pricing   # the page as Markdown, chrome removed
aglyn llms example.com                   # what the site is for, and which URL answers what
aglyn openapi example.com                # its OpenAPI 3.1 description
aglyn site example.com                   # its public identity, as JSON
aglyn pages example.com                  # every published page, following the cursor
aglyn sitemap example.com                # the sitemap index

aglyn read is the one to reach for from a script or an agent. It asks the site for text/markdown, so what comes back is the prose — no navigation, no styling, no <div> soup to parse and guess at.

aglyn read https://example.com/blog/hello | wc -w
aglyn pages example.com --json | jq -r '.[].slug'

The REST API

aglyn api is the only command that authenticates. Create a key in the console under Settings → API keys, then:

export AGLYN_API_KEY=aglyn_sk_...
aglyn api GET /sites
aglyn auth status          # says whether a key is set; never prints it

The CLI does not write your key to disk. It reads AGLYN_API_KEY from the environment so the key lives wherever you already keep secrets.

Exit codes

| code | meaning | | ---- | ------- | | 0 | it worked (--help included) | | 1 | the request failed, or the site answered an error | | 2 | the command was used wrongly |

Options

| flag | effect | | ---- | ------ | | --json | JSON instead of a table, where both make sense | | --limit <n> | page size for list commands (default 50, max 100) | | --base <url> | API base for api (default AGLYN_API_URL) | | -h, --help | usage | | -V, --version | version |

Licence

Apache-2.0