diff --git a/apps/orb_trader/engine.py b/apps/orb_trader/engine.py
index 9f4413d..81f02bf 100644
--- a/apps/orb_trader/engine.py
+++ b/apps/orb_trader/engine.py
@@ -109,12 +109,6 @@ class ORBTradingEngine:
                 raw_bars.update(self._broker.get_bars(chunk, start, today))
             except Exception as e:
                 self._log(f"  WARNING: daily bars chunk {i//chunk_size+1} failed ({e}) — skipping")
-                # Retry chunk symbol-by-symbol to isolate the bad ticker(s)
-                for sym in chunk:
-                    try:
-                        raw_bars.update(self._broker.get_bars([sym], start, today))
-                    except Exception:
-                        self._log(f"  Skipping invalid symbol: {sym}")
 
         daily_bars_dict = bars_to_enrichment_format(raw_bars)
 
diff --git a/apps/paper_trader/alpaca_broker.py b/apps/paper_trader/alpaca_broker.py
index abcf61c..773b3c0 100644
--- a/apps/paper_trader/alpaca_broker.py
+++ b/apps/paper_trader/alpaca_broker.py
@@ -225,35 +225,33 @@ class AlpacaBroker:
         start: dt.date,
         end: dt.date,
     ) -> dict[str, list[Bar]]:
-        """Fetch daily OHLCV bars for a list of symbols in [start, end]."""
+        """Fetch daily OHLCV bars for a list of symbols via Oracle API.
+
+        Oracle normalises problematic symbols (e.g. BF-B → BF.B) and maps
+        responses back to the original symbol names.
+        """
         if not symbols:
             return {}
 
-        from alpaca.data.requests import StockBarsRequest
-        from alpaca.data.timeframe import TimeFrame
+        from libs.oracle_client.alpaca import get_multi_daily_bars
 
-        req = StockBarsRequest(
-            symbol_or_symbols=symbols,
-            timeframe=TimeFrame.Day,
-            start=dt.datetime.combine(start, dt.time.min),
-            end=dt.datetime.combine(end, dt.time.max),
-            feed="iex",
+        raw = get_multi_daily_bars(
+            tickers=symbols,
+            start_date=start.isoformat(),
+            end_date=end.isoformat(),
         )
-        response = self._data.get_stock_bars(req)
+
         result: dict[str, list[Bar]] = {}
         for sym in symbols:
-            try:
-                bars_data = response[sym]
-            except (KeyError, TypeError):
-                bars_data = []
+            bars_data = raw.get(sym, [])
             result[sym] = [
                 Bar(
-                    date=b.timestamp.date().isoformat() if hasattr(b.timestamp, "date") else str(b.timestamp)[:10],
-                    open=float(b.open),
-                    high=float(b.high),
-                    low=float(b.low),
-                    close=float(b.close),
-                    volume=float(b.volume),
+                    date=b["date"],
+                    open=float(b["open"]),
+                    high=float(b["high"]),
+                    low=float(b["low"]),
+                    close=float(b["close"]),
+                    volume=float(b["volume"]),
                 )
                 for b in bars_data
             ]
@@ -283,6 +281,35 @@ class AlpacaBroker:
             result[sym] = date_map
         return result
 
+    def get_intraday_bars(
+        self,
+        symbols: list[str],
+        start: dt.datetime,
+        end: dt.datetime,
+        timeframe_minutes: int = 5,
+    ) -> dict[str, list[dict]]:
+        """Fetch intraday OHLCV bars for a list of symbols via Oracle API.
+
+        Returns {symbol: [{timestamp: ISO8601, open, high, low, close, volume}, ...]}.
+        """
+        if not symbols:
+            return {}
+
+        from libs.oracle_client.alpaca import get_multi_intraday_bars
+
+        interval = f"{timeframe_minutes}min"
+        raw = get_multi_intraday_bars(
+            tickers=symbols,
+            start_date=start.date().isoformat(),
+            end_date=end.date().isoformat(),
+            interval=interval,
+        )
+
+        result: dict[str, list[dict]] = {sym: [] for sym in symbols}
+        for sym in symbols:
+            result[sym] = raw.get(sym, [])
+        return result
+
     def get_latest_bars(self, symbols: list[str]) -> dict[str, Bar]:
         """Fetch the latest bar for each symbol."""
         if not symbols:
diff --git a/libs/oracle_client/alpaca.py b/libs/oracle_client/alpaca.py
new file mode 100644
index 0000000..5454ffd
--- /dev/null
+++ b/libs/oracle_client/alpaca.py
@@ -0,0 +1,197 @@
+"""Alpaca real-time snapshot service via Stock Oracle API.
+
+Provides synchronous helpers for real-time price snapshots, suitable for use
+in threaded engine code (which already runs in asyncio.to_thread).
+"""
+from __future__ import annotations
+
+import logging
+import os
+from dataclasses import dataclass
+from typing import Any
+
+log = logging.getLogger(__name__)
+
+
+@dataclass
+class AlpacaSnapshot:
+    """Real-time Alpaca snapshot for a single ticker."""
+
+    ticker: str
+    price: float | None = None         # last trade price
+    bid: float | None = None
+    ask: float | None = None
+    bid_size: int | None = None
+    ask_size: int | None = None
+    open: float | None = None
+    high: float | None = None
+    low: float | None = None
+    volume: int | None = None
+    vwap: float | None = None
+    prev_close: float | None = None
+    change: float | None = None
+    change_pct: float | None = None
+    timestamp: str | None = None
+
+    @property
+    def mid(self) -> float | None:
+        """Midpoint between bid and ask, or last trade if unavailable."""
+        if self.bid is not None and self.ask is not None:
+            return (self.bid + self.ask) / 2
+        return self.price
+
+
+def _base_url() -> str:
+    return os.environ.get("ORACLE_URL", "http://localhost:18001").rstrip("/")
+
+
+def _parse_snapshot(data: dict[str, Any]) -> AlpacaSnapshot:
+    return AlpacaSnapshot(
+        ticker=data["ticker"],
+        price=data.get("price"),
+        bid=data.get("bid"),
+        ask=data.get("ask"),
+        bid_size=data.get("bid_size"),
+        ask_size=data.get("ask_size"),
+        open=data.get("open"),
+        high=data.get("high"),
+        low=data.get("low"),
+        volume=data.get("volume"),
+        vwap=data.get("vwap"),
+        prev_close=data.get("prev_close"),
+        change=data.get("change"),
+        change_pct=data.get("change_pct"),
+        timestamp=data.get("timestamp"),
+    )
+
+
+def get_multi_daily_bars(
+    tickers: list[str],
+    start_date: str,
+    end_date: str,
+    base_url: str | None = None,
+) -> dict[str, list[dict]]:
+    """Fetch daily OHLCV bars for multiple tickers via Oracle API.
+
+    Oracle handles symbol normalisation (e.g. BF-B → BF.B) and Alpaca batching
+    internally.  Returns {ticker: [{date, open, high, low, close, volume}, ...]}.
+    Missing/errored tickers are omitted.
+    """
+    import httpx
+
+    if not tickers:
+        return {}
+
+    result: dict[str, list[dict]] = {}
+    url = (base_url or _base_url()) + "/api/v1/price/data"
+
+    # Oracle handles its own batching, but we chunk here as a safety net for
+    # very large URL query strings.
+    chunk_size = 300
+    for i in range(0, len(tickers), chunk_size):
+        chunk = tickers[i : i + chunk_size]
+        try:
+            resp = httpx.get(
+                url,
+                params={"tickers": ",".join(chunk), "start_date": start_date, "end_date": end_date},
+                timeout=90.0,
+            )
+            resp.raise_for_status()
+            for ticker, bars in resp.json().get("bars", {}).items():
+                result[ticker] = bars
+        except Exception as exc:
+            log.warning("Oracle multi_daily_bars chunk %d failed: %s", i // chunk_size, exc)
+
+    return result
+
+
+def get_multi_intraday_bars(
+    tickers: list[str],
+    start_date: str,
+    end_date: str,
+    interval: str = "5min",
+    base_url: str | None = None,
+) -> dict[str, list[dict]]:
+    """Fetch intraday bars for multiple tickers via Oracle API.
+
+    Returns {ticker: [{timestamp (ISO8601), open, high, low, close, volume}, ...]}.
+    """
+    import httpx
+
+    if not tickers:
+        return {}
+
+    result: dict[str, list[dict]] = {}
+    url = (base_url or _base_url()) + "/api/v1/alpaca/intraday"
+
+    chunk_size = 300
+    for i in range(0, len(tickers), chunk_size):
+        chunk = tickers[i : i + chunk_size]
+        try:
+            resp = httpx.get(
+                url,
+                params={
+                    "tickers": ",".join(chunk),
+                    "interval": interval,
+                    "start_date": start_date,
+                    "end_date": end_date,
+                },
+                timeout=90.0,
+            )
+            resp.raise_for_status()
+            for ticker, bars in resp.json().get("bars", {}).items():
+                result[ticker] = bars
+        except Exception as exc:
+            log.warning("Oracle multi_intraday_bars chunk %d failed: %s", i // chunk_size, exc)
+
+    return result
+
+
+def get_snapshot(ticker: str, base_url: str | None = None) -> AlpacaSnapshot | None:
+    """Fetch real-time snapshot for a single ticker (synchronous).
+
+    Returns None on any error.
+    """
+    import httpx
+
+    url = (base_url or _base_url()) + f"/api/v1/alpaca/snapshot/{ticker}"
+    try:
+        resp = httpx.get(url, timeout=10.0)
+        resp.raise_for_status()
+        return _parse_snapshot(resp.json())
+    except Exception as exc:
+        log.warning("Oracle snapshot(%s) failed: %s", ticker, exc)
+        return None
+
+
+def get_snapshots(
+    tickers: list[str],
+    base_url: str | None = None,
+    batch_size: int = 200,
+) -> dict[str, AlpacaSnapshot]:
+    """Fetch real-time snapshots for multiple tickers (synchronous).
+
+    Batches requests and returns {ticker: AlpacaSnapshot}.
+    Missing/errored tickers are omitted.
+    """
+    import httpx
+
+    if not tickers:
+        return {}
+
+    result: dict[str, AlpacaSnapshot] = {}
+    url = (base_url or _base_url()) + "/api/v1/alpaca/snapshot"
+
+    for i in range(0, len(tickers), batch_size):
+        chunk = tickers[i : i + batch_size]
+        try:
+            resp = httpx.get(url, params={"tickers": ",".join(chunk)}, timeout=15.0)
+            resp.raise_for_status()
+            data = resp.json()
+            for snap_data in data.get("snapshots", []):
+                snap = _parse_snapshot(snap_data)
+                result[snap.ticker] = snap
+        except Exception as exc:
+            log.warning("Oracle snapshots(batch %d) failed: %s", i // batch_size, exc)
+
+    return result