diff --git a/plugins/platforms/zulip/adapter.py b/plugins/platforms/zulip/adapter.py
index a9c6877..125a807 100644
--- a/plugins/platforms/zulip/adapter.py
+++ b/plugins/platforms/zulip/adapter.py
@@ -43,6 +43,7 @@ import time
 import uuid
 from collections import deque
 from datetime import datetime, timezone
+from pathlib import Path
 from typing import Any, Dict, List, Optional, Tuple
 
 try:
@@ -881,6 +882,116 @@ class ZulipAdapter(BasePlatformAdapter):
                 reply_to_topic=topic,
             )
 
+    # ------------------------------------------------------------------
+    # Media / attachment extraction (vision support)
+    # ------------------------------------------------------------------
+
+    async def _extract_inline_media(
+        self, content: str, media_urls: List[str], media_types: List[str],
+    ) -> None:
+        """Download Zulip ``/user_uploads/...`` attachments referenced in a
+        message so vision/transcription tools can read them locally.
+
+        Zulip renders uploaded files as ``<img src="/user_uploads/...">``
+        in HTML. After ``_strip_html()`` the markdown-style references
+        survive in the plain text. This method finds those references,
+        downloads the files, and caches them into Hermes media storage.
+        """
+        for rel in re.findall(r"\]\((/user_uploads/[^)]+)\)", content):
+            dl_url = f"{self._site}{rel}"
+            try:
+                resp = await self._http_client.get(
+                    dl_url,
+                    headers={"Authorization": self._auth_header},
+                    timeout=30.0,
+                )
+                if resp.status_code >= 400:
+                    logger.debug(
+                        "[%s] Inline media %s -> %d",
+                        self.name, rel, resp.status_code,
+                    )
+                    continue
+                data = resp.content
+                mime = (
+                    resp.headers.get("content-type", "application/octet-stream")
+                    .split(";")[0]
+                    .strip()
+                )
+            except Exception:
+                logger.debug("[%s] Failed to download inline media %s", self.name, rel)
+                continue
+            fname = rel.rsplit("/", 1)[-1] or "file"
+            await self._cache_attachment(data, fname, mime, media_urls, media_types)
+
+    async def _extract_event_attachments(
+        self, message: Dict[str, Any], media_urls: List[str], media_types: List[str],
+    ) -> None:
+        """Download files from Zulip's ``message.attachments`` field (UI file
+        uploads that Zulip sends as structured event data rather than inline
+        markdown links)."""
+        atts = message.get("attachments") or []
+        for att in atts:
+            path_id = att.get("path_id")
+            fname = att.get("name") or "file"
+            if not path_id:
+                continue
+            dl_url = f"{self._site}/api/v1/user_uploads/{path_id}"
+            try:
+                resp = await self._http_client.get(
+                    dl_url,
+                    headers={"Authorization": self._auth_header},
+                    timeout=30.0,
+                )
+                if resp.status_code >= 400:
+                    logger.debug(
+                        "[%s] Event attachment %s -> %d",
+                        self.name, fname, resp.status_code,
+                    )
+                    continue
+                data = resp.content
+                mime = (
+                    resp.headers.get("content-type", "application/octet-stream")
+                    .split(";")[0]
+                    .strip()
+                )
+            except Exception:
+                logger.debug("[%s] Failed to download event attachment %s", self.name, fname)
+                continue
+            await self._cache_attachment(data, fname, mime, media_urls, media_types)
+
+    async def _cache_attachment(
+        self, data: bytes, fname: str, mime: str,
+        media_urls: List[str], media_types: List[str],
+    ) -> None:
+        """Cache a downloaded attachment into Hermes media storage so the
+        vision_analyze / ocr tools can access it via a local file path."""
+        try:
+            from gateway.platforms.base import (
+                cache_image_from_bytes,
+                cache_audio_from_bytes,
+                cache_document_from_bytes,
+            )
+        except ImportError:
+            logger.warning("[%s] cache_*_from_bytes not available in gateway", self.name)
+            return
+        ext = Path(fname).suffix
+        try:
+            if mime.startswith("image/"):
+                media_urls.append(cache_image_from_bytes(data, ext or ".png"))
+                media_types.append(mime)
+            elif mime.startswith("audio/"):
+                media_urls.append(cache_audio_from_bytes(data, ext or ".ogg"))
+                media_types.append(mime)
+            else:
+                media_urls.append(cache_document_from_bytes(data, fname))
+                media_types.append(mime)
+        except Exception as exc:
+            logger.warning("[%s] Failed to cache attachment %s: %s", self.name, fname, exc)
+
+    # ------------------------------------------------------------------
+    # Message routing
+    # ------------------------------------------------------------------
+
     async def _route_message(
         self,
         text: str,
@@ -911,8 +1022,22 @@ class ZulipAdapter(BasePlatformAdapter):
             chat_id_alt=str(reply_to_id) if reply_to_id else None,
         )
 
-        # Derive message type
-        message_type = MessageType.TEXT
+        # Gen 5: Extract media/attachments so vision tools can read them
+        media_urls: List[str] = []
+        media_types: List[str] = []
+        await self._extract_inline_media(text, media_urls, media_types)
+        await self._extract_event_attachments(raw, media_urls, media_types)
+
+        # Derive message type — upgrade to PHOTO/VOICe/DOCUMENT when media present
+        if media_types:
+            if any(m.startswith("image/") for m in media_types):
+                message_type = MessageType.PHOTO
+            elif any(m.startswith("audio/") for m in media_types):
+                message_type = MessageType.VOICE
+            else:
+                message_type = MessageType.DOCUMENT
+        else:
+            message_type = MessageType.TEXT
 
         message_event = MessageEvent(
             text=text,
@@ -921,6 +1046,8 @@ class ZulipAdapter(BasePlatformAdapter):
             message_id=msg_id,
             raw_message=raw,
             timestamp=_parse_zulip_timestamp(timestamp),
+            media_urls=media_urls or None,
+            media_types=media_types or None,
         )
 
         # Store reply metadata so send() can respond to the right thread