docs: every install command site-wide is now a real package

Four surfaces still told users to 'pip install secryst' or
'npm i secryst' — packages that are not ours (Python runtime is
interscript-ml; the npm package is interscript). The phonological
layer page, about page, ml page, and the Secryst usage doc are
corrected; the usage doc is rewritten in plain language with working
examples. Zero wrong install commands remain under src/.

Author: ronaldtse

src/content/docs/Usage_with_Secryst.adoc

@@ -1,36 +1,48 @@
 = Usage with Secryst
 
-Secryst is a seq2seq transformer suited for transliteration. Written in Ruby.
-It's installation is a bit tricky, you should consult its own installation guide
-(https://github.com/secryst/secryst[at GitHub]). By default we don't use Secryst,
-unless you have installed it.
+Secryst is the Ruby runtime for Interscript's neural models:
 
-== Using it standalone
+[source,ruby]
+----
+gem install secryst
+----
 
-It's enough to install it. Be sure to consult the guide above.
+For details, see the https://github.com/secryst/secryst[repository].
 
-== Integration with Ruby Applications
+== Using it standalone
 
-In your Gemfile, add:
+After installing, load any published model by id:
 
 [source,ruby]
 ----
-source "https://rubygems.org"
+require "secryst"
 
-gem "secryst"
+model = Secryst::Model.load("ara-diac-small-2.1-int8")
+model.translate("قوله فحكمها في الوفاة")
 ----
 
-Create a Secrystfile near your Gemfile with the following, for each model you
-want to use in your application. Please consult our Secrystfile to get all the
-maps needed to get all the Secryst maps needed.
+The model is downloaded once, verified against its checksum, and cached
+under `~/.cache/secryst`. The list of published models is in
+https://github.com/interscript/interscript-ml/blob/main/models.yaml[models.yaml].
+
+== Integration with Ruby applications
+
+Add the gem to your Gemfile:
 
 [source,ruby]
 ----
-model "model-name"
+source "https://rubygems.org"
+
+gem "secryst"
 ----
 
+Load models where they are needed. `Model.load` resolves, verifies, and
+caches on first use; there is no separate configuration file.
+
 == Usage inside maps
 
+A transliteration map can run a neural model as a step:
+
 [source,ruby]
 ----
 stage {
@@ -40,4 +52,5 @@ stage {
 }
 ----
 
-As of now, Secryst is usable only by the Ruby implementation.
+A map that declares a `secryst` step requires the gem; maps without
+neural steps work unchanged.

src/pages/about.astro

@@ -172,8 +172,8 @@ const partners: Partner[] = [
           </li>
           <li>
             <strong>Crystals</strong> — <a href="https://www.secryst.org">secryst</a> (<code
-              >gem secryst</code
-            > · <code>pip install secryst</code> · <code>npm i secryst</code>) — local ONNX
+              >gem install secryst</code
+            > · <code>pip install interscript-ml</code> · <code>npm install interscript</code>) — local ONNX
             inference implementing the contract. A crystal has no interscript dependency; an engine
             without a crystal simply cannot execute maps that declare a vocalization step.
           </li>

src/pages/docs/phonological-layer.astro

@@ -32,7 +32,7 @@ import Base from "../../layouts/Base.astro"
           <li><strong>interscript engines</strong> stay pure. A map that needs
             vocalization declares a <code>secryst</code> (or future <code>vocalize</code>) step; without a crystal installed, that map
             simply cannot run that step.</li>
-          <li><strong>secryst crystals</strong> (<code>gem secryst</code> · <code>pip install secryst</code> · <code>npm i secryst</code> resolve a model id through <code>models.yaml</code>, download a sha256-verified IMF&nbsp;v1 zip,
+          <li><strong>secryst crystals</strong> (<code>gem install secryst</code> · <code>pip install interscript-ml</code> · <code>npm install interscript</code> resolve a model id through <code>models.yaml</code>, download a sha256-verified IMF&nbsp;v1 zip,
             cache it under <code>~/.cache/secryst</code>, and run greedy ONNX
             inference locally. No cloud, no telemetry.</li>
           <li><strong>interscript-ml</strong> is the contract only: the index,

src/pages/ml.astro

@@ -454,7 +454,7 @@ const clientModels = [
 gem install secryst
 
 # Python
-pip install secryst
+pip install interscript-ml
 
 # TypeScript / JavaScript
 npm install secryst</code></pre>