Chat mode, phase 1: grounded streaming chat over your notes

Entry points (no sixth tab): Ask in the Library header (desktop button,
phone icon), Ask about this deck in the deck menu, Ask a question on
study/homework results (compact result travels as context), and a Chat
row in the desktop sidebar. Phones get a full-screen route with back +
New chat; the tab bar steps aside.

- Client-side retrieval: chunk scoped decks (~800 tokens), BM25 against
  the question + recent turns, top chunks under ~7k tokens, each
  labelled with deck + section. Notes come from the deck's stored
  sourceText (already kept for Regenerate); decks without it fall back
  to generated content and the UI says so.
- Stateless server: /api/chat validates against the CHAT_MODELS
  allowlist, trims the history to ~6k tokens, caps context at ~9k and
  streams token/thinking/done/error NDJSON. /api/chat/extract reduces
  attachments to text through the existing vision step.
- UI: user bubbles + full-width assistant text, streaming with stop,
  auto-scroll that pauses with a jump-to-bottom button, safe markdown
  (React nodes, no HTML injection; KaTeX dynamically imported only when
  a message has math), Thinking row for reasoning, copy/regenerate/
  edit-and-resend, model + context pickers (last model remembered),
  composer with visualViewport keyboard handling, safe-area, 16px
  textarea, Enter-to-send on desktop.
- Usage-limit failures keep the draft and show the shared message with
  Settings/Ollama actions; stopping maps the reader abort to a clean
  stopped state; offline disables sending with a reason.
- Settings: Keep notes text for chat toggle; deck menu: Delete notes
  text. RATE_LIMIT_CHAT + CHAT_MODELS in .env.example and README.
- Tests: chat suite (63 checks, stub Ollama) wired into run-all.
  Full set: 12/12 suites, 487 checks. SW -> freezmo-v32.

Also: tests/model-compare.mjs (diagnostic) benchmarks two models on the
same notes (latency, repair rounds, activity types) so the gpt-oss vs
gemma default can be decided on numbers; defaults unchanged.

lib/prompts.ts

@@ -367,6 +367,36 @@ export function buildRefineUserPrompt(input: {
     .join("\n");
 }
 
+/* ------------------------------------------------------------------ */
+/* Chat                                                                */
+/* ------------------------------------------------------------------ */
+
+export const CHAT_SYSTEM_PROMPT = `You are Freezmo's chat tutor. You talk with a student about their own study notes and help them understand, review and quiz themselves.
+
+Rules:
+1. The attached notes are the primary source. Ground your answer in them and prefer them over general knowledge.
+2. If the notes do not cover something, say so plainly first ("Your notes don't cover this"), then answer from general knowledge and label that part clearly as general knowledge — never blur the two.
+3. Never claim the notes say something they do not, and never attribute invented facts to the notes.
+4. Answer in the language the student writes in (not necessarily the app's language), unless they ask for another language.
+5. Be a concise, friendly teacher: short paragraphs, lists when they help, no filler, no long preambles. If the student asks to be quizzed, ask one question at a time and wait for their answer before checking it.
+6. Markdown is supported (lists, tables, \`code\`, formulas as $...$). Keep tables narrow enough for a phone.`;
+
+export function buildChatSystemPrompt(input: {
+  context: { deckTitle: string; section: string; text: string }[];
+}): string {
+  if (input.context.length === 0) {
+    return [
+      CHAT_SYSTEM_PROMPT,
+      "",
+      "No notes are attached to this conversation: the student chose general chat. Answer from general knowledge and, when it matters, remind them they can scope the chat to their notes.",
+    ].join("\n");
+  }
+  const blocks = input.context
+    .map((chunk) => `### ${chunk.deckTitle} — ${chunk.section}\n${chunk.text}`)
+    .join("\n\n");
+  return [CHAT_SYSTEM_PROMPT, "", "The student's notes (primary source):", "", blocks].join("\n");
+}
+
 /* ------------------------------------------------------------------ */
 /* Retry helper                                                        */
 /* ------------------------------------------------------------------ */