docs: restructure README for the v0.15 built-in-engine launch

docs/agentic-search.md

@@ -8,7 +8,7 @@ A deep dive into Thuki's agentic RAG search pipeline: how retrieval-augmented ge
 
 ## Setup
 
-> **Roadmap:** First-class, out-of-box `/search` support (bundled native sidecars or pre-built container images shipped with the app) is planned. Today, enabling `/search` requires cloning this repository to run the local Docker services described below. Track progress and contribute in the project's [GitHub issues](https://github.com/quiet-node/thuki/issues).
+> **Roadmap:** First-class, out-of-box `/search` support (bundled native sidecars or pre-built container images shipped with the app) is planned. Today, enabling `/search` requires cloning this repository to run the local Docker services described below. Track it in [#257](https://github.com/quiet-node/thuki/issues/257).
 
 The `/search` command depends on two local Docker containers: a **SearXNG** meta-search engine and a **Trafilatura** reader. Both run on `127.0.0.1` only, so every query and every fetched page stays on your machine.
 

docs/commands.md

@@ -18,7 +18,7 @@ This means you can highlight text anywhere on screen, summon Thuki with double-t
 
 ## Image input on text-only models
 
-`/extract`, `/tldr`, `/translate`, `/rewrite`, `/refine`, `/bullets`, `/todos`, and `/explain` read attached images locally via macOS Vision OCR, so they work even when the active model has no vision capability. Only plain submits and `/screen` alone require a vision model to read images. See [OCR-supported commands](./ocr-commands.md) for the full list and details.
+`/extract`, `/tldr`, `/translate`, `/rewrite`, `/refine`, `/bullets`, `/todos`, and `/explain` pull the text out of attached images locally via macOS Vision OCR, so they work even when the active model has no vision capability. They read text, not scenes: describing a textless image still needs a vision model. Only plain submits and `/screen` alone send the image itself to the model. See [OCR-supported commands](./ocr-commands.md) for the full list and details.
 
 ## /search
 

sandbox/search-box/reader/README.md

@@ -2,6 +2,8 @@
 
 Trafilatura-based URL-to-markdown extractor. Second stop of Thuki's agentic `/search` pipeline.
 
+> **Roadmap:** This Docker service is the interim way to run `/search`. A first-class, out-of-box `/search` bundled with the app (no Docker to run yourself) is planned. See [Setup](../../../docs/agentic-search.md#setup).
+
 ## What it does
 
 Takes a URL, fetches the page, strips boilerplate (navigation, ads, footers, cookie banners), and returns clean markdown the synthesis LLM can cite against.

sandbox/search-box/searxng/README.md

@@ -2,6 +2,8 @@
 
 Self-hosted, privacy-respecting meta-search engine. First stop of Thuki's `/search` pipeline.
 
+> **Roadmap:** This Docker service is the interim way to run `/search`. A first-class, out-of-box `/search` bundled with the app (no Docker to run yourself) is planned. See [Setup](../../../docs/agentic-search.md#setup).
+
 ## What it does
 
 SearXNG aggregates results from 70+ upstream engines (Google, Bing, DuckDuckGo, Brave, Wikipedia, Stack Overflow, GitHub, arXiv, etc.) and returns a unified JSON result set. Our instance is:

src/config/commandArtifacts.ts

@@ -69,7 +69,7 @@ export function renderCommandsMarkdown(): string {
     '',
     '## Image input on text-only models',
     '',
-    '`/extract`, `/tldr`, `/translate`, `/rewrite`, `/refine`, `/bullets`, `/todos`, and `/explain` read attached images locally via macOS Vision OCR, so they work even when the active model has no vision capability. Only plain submits and `/screen` alone require a vision model to read images. See [OCR-supported commands](./ocr-commands.md) for the full list and details.',
+    '`/extract`, `/tldr`, `/translate`, `/rewrite`, `/refine`, `/bullets`, `/todos`, and `/explain` pull the text out of attached images locally via macOS Vision OCR, so they work even when the active model has no vision capability. They read text, not scenes: describing a textless image still needs a vision model. Only plain submits and `/screen` alone send the image itself to the model. See [OCR-supported commands](./ocr-commands.md) for the full list and details.',
     '',
     ...COMMANDS.flatMap((command, index) => {
       const section = renderDocsSection(command);

src/config/tips.ts

@@ -11,7 +11,7 @@ export const TIPS: readonly Tip[] = [
   'Use /screen to snap your display and attach it to the chat for visual context',
   'Use /extract to extract any text from images or screenshots',
   {
-    text: 'OCR-supported commands (/extract, /translate, etc.) read images locally: no vision model needed ↗',
+    text: 'OCR-supported commands (/extract, /translate, etc.) read text in images locally: no vision model needed ↗',
     url: 'https://github.com/quiet-node/thuki/blob/main/docs/ocr-commands.md',
   },
   'Highlight text in any app before summoning Thuki to include it as context',

src/view/onboarding/SubscribeStep.tsx

@@ -230,9 +230,10 @@ export function SubscribeStep({ onContinue, downloadStatus }: Props) {
           ))}
         </div>
 
-        {/* Free + local guarantee, capping the roadmap. It applies to the list
-            specifically (voice and meeting notes are paid, cloud features
-            elsewhere) and is the only product-local line on this screen, so it
+        {/* Local-first ethos line, capping the roadmap. It states how Thuki is
+            built today rather than promising a delivery model for the unshipped
+            roadmap items above, whose pricing and where they run are not yet
+            decided. It is the only product-local line on this screen, so it
             repeats neither the subtitle nor the email trust line below. */}
         <p
           style={{
@@ -243,7 +244,7 @@ export function SubscribeStep({ onContinue, downloadStatus }: Props) {
             margin: '0 0 12px',
           }}
         >
-          All free. All local. Nothing ever leaves your Mac.
+          Local-first, always. Your conversations stay on your Mac.
         </p>
 
         {/* Divider */}

src/view/onboarding/__tests__/SubscribeStep.test.tsx

@@ -41,10 +41,12 @@ describe('SubscribeStep', () => {
     ).toBeInTheDocument();
   });
 
-  it('states the free and local guarantee under the roadmap', () => {
+  it('states the local-first ethos line under the roadmap', () => {
     render(<SubscribeStep onContinue={vi.fn()} />);
     expect(
-      screen.getByText('All free. All local. Nothing ever leaves your Mac.'),
+      screen.getByText(
+        'Local-first, always. Your conversations stay on your Mac.',
+      ),
     ).toBeInTheDocument();
   });