Initial commit: webmail client (M0/M1)

Three-pane webmail client.
- Backend: FastAPI + async SQLAlchemy + arq worker (Postgres/Redis), IMAP/SMTP
  sync, threading, search, drafts/outbox with undo, attachments, 2FA.
- Frontend: React 18 + Vite + TypeScript + TanStack Query + Tailwind; three-pane
  UI with rich-text composer, light/dark themes, resizable panels.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

backend/app/api/recipients.py

@@ -0,0 +1,52 @@
+"""Recipient autocomplete from known_recipients (SPEC §9.1, §4).
+
+Prefix match on email or name, ranked by ``times_used`` decayed by recency. The
+composer merges these with plugin capability providers client-side (M3).
+"""
+
+from __future__ import annotations
+
+from typing import Annotated
+
+from fastapi import APIRouter, Depends, Query
+from pydantic import BaseModel
+from sqlalchemy import func, select
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.db import get_session
+from app.identity.deps import CurrentUser
+from app.models import KnownRecipient
+
+router = APIRouter(prefix="/recipients", tags=["compose"])
+
+DbSession = Annotated[AsyncSession, Depends(get_session)]
+
+
+class Suggestion(BaseModel):
+    name: str
+    email: str
+
+
+@router.get("/suggest", response_model=list[Suggestion])
+async def suggest(
+    user: CurrentUser, db: DbSession, q: str = Query("", max_length=200)
+) -> list[Suggestion]:
+    term = q.strip()
+    if not term:
+        return []
+    like = f"{term}%"
+    # Recency-decayed score: times_used / (1 + age_in_days). Ranks frequent and
+    # recent correspondents ahead of stale ones (SPEC §9.1).
+    age_days = func.extract("epoch", func.now() - KnownRecipient.last_used_at) / 86400.0
+    score = KnownRecipient.times_used / (1.0 + age_days)
+    stmt = (
+        select(KnownRecipient.name, KnownRecipient.email)
+        .where(
+            KnownRecipient.user_id == user.id,
+            KnownRecipient.email.ilike(like) | KnownRecipient.name.ilike(like),
+        )
+        .order_by(score.desc())
+        .limit(10)
+    )
+    rows = (await db.execute(stmt)).all()
+    return [Suggestion(name=row.name, email=row.email) for row in rows]