Switch image analysis from Ollama to llama-server

Use the same llama-server (OpenAI-compatible API) for vision analysis
that the main agent uses, with OpenAI fallback. Sends images as base64
in the standard OpenAI vision message format.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

app.py

@@ -13,6 +13,7 @@ import blueprints.users
 import blueprints.whatsapp
 import blueprints.email
 import blueprints.users.models
+from config.db import TORTOISE_CONFIG
 from main import consult_simba_oracle
 
 # Load environment variables
@@ -28,6 +29,7 @@ app = Quart(
 )
 
 app.config["JWT_SECRET_KEY"] = os.getenv("JWT_SECRET_KEY", "SECRET_KEY")
+app.config["MAX_CONTENT_LENGTH"] = 10 * 1024 * 1024  # 10 MB upload limit
 jwt = JWTManager(app)
 
 # Register blueprints
@@ -38,24 +40,6 @@ app.register_blueprint(blueprints.whatsapp.whatsapp_blueprint)
 app.register_blueprint(blueprints.email.email_blueprint)
 
 
-# Database configuration with environment variable support
-DATABASE_URL = os.getenv(
-    "DATABASE_URL", "postgres://raggr:raggr_dev_password@localhost:5432/raggr"
-)
-
-TORTOISE_CONFIG = {
-    "connections": {"default": DATABASE_URL},
-    "apps": {
-        "models": {
-            "models": [
-                "blueprints.conversation.models",
-                "blueprints.users.models",
-                "aerich.models",
-            ]
-        },
-    },
-}
-
 # Initialize Tortoise ORM with lifecycle hooks
 @app.while_serving
 async def lifespan():

blueprints/conversation/__init__.py

@@ -1,13 +1,19 @@
 import datetime
 import json
+import logging
+import uuid
 
-from quart import Blueprint, jsonify, make_response, request
+from quart import Blueprint, Response, jsonify, make_response, request
 from quart_jwt_extended import (
     get_jwt_identity,
     jwt_refresh_token_required,
 )
 
 import blueprints.users.models
+from utils.image_process import analyze_user_image
+from utils.image_upload import ImageValidationError, process_image
+from utils.s3_client import get_image as s3_get_image
+from utils.s3_client import upload_image as s3_upload_image
 
 from .agents import main_agent
 from .logic import (
@@ -29,7 +35,9 @@ conversation_blueprint = Blueprint(
 _SYSTEM_PROMPT = SIMBA_SYSTEM_PROMPT
 
 
-def _build_messages_payload(conversation, query_text: str) -> list:
+def _build_messages_payload(
+    conversation, query_text: str, image_description: str | None = None
+) -> list:
     recent_messages = (
         conversation.messages[-10:]
         if len(conversation.messages) > 10
@@ -38,8 +46,19 @@ def _build_messages_payload(conversation, query_text: str) -> list:
     messages_payload = [{"role": "system", "content": _SYSTEM_PROMPT}]
     for msg in recent_messages[:-1]:  # Exclude the message we just added
         role = "user" if msg.speaker == "user" else "assistant"
-        messages_payload.append({"role": role, "content": msg.text})
-    messages_payload.append({"role": "user", "content": query_text})
+        text = msg.text
+        if msg.image_key and role == "user":
+            text = f"[User sent an image]\n{text}"
+        messages_payload.append({"role": role, "content": text})
+
+    # Build the current user message with optional image description
+    if image_description:
+        content = f"[Image analysis: {image_description}]"
+        if query_text:
+            content = f"{query_text}\n\n{content}"
+    else:
+        content = query_text
+    messages_payload.append({"role": "user", "content": content})
     return messages_payload
 
 
@@ -74,6 +93,58 @@ async def query():
     return jsonify({"response": message})
 
 
+@conversation_blueprint.post("/upload-image")
+@jwt_refresh_token_required
+async def upload_image():
+    current_user_uuid = get_jwt_identity()
+    await blueprints.users.models.User.get(id=current_user_uuid)
+
+    files = await request.files
+    form = await request.form
+    file = files.get("file")
+    conversation_id = form.get("conversation_id")
+
+    if not file or not conversation_id:
+        return jsonify({"error": "file and conversation_id are required"}), 400
+
+    file_bytes = file.read()
+    content_type = file.content_type or "image/jpeg"
+
+    try:
+        processed_bytes, output_content_type = process_image(file_bytes, content_type)
+    except ImageValidationError as e:
+        return jsonify({"error": str(e)}), 400
+
+    ext = output_content_type.split("/")[-1]
+    if ext == "jpeg":
+        ext = "jpg"
+    key = f"conversations/{conversation_id}/{uuid.uuid4()}.{ext}"
+
+    await s3_upload_image(processed_bytes, key, output_content_type)
+
+    return jsonify(
+        {
+            "image_key": key,
+            "image_url": f"/api/conversation/image/{key}",
+        }
+    )
+
+
+@conversation_blueprint.get("/image/<path:image_key>")
+@jwt_refresh_token_required
+async def serve_image(image_key: str):
+    try:
+        image_bytes, content_type = await s3_get_image(image_key)
+    except Exception:
+        return jsonify({"error": "Image not found"}), 404
+
+    return Response(
+        image_bytes,
+        content_type=content_type,
+        headers={"Cache-Control": "private, max-age=3600"},
+    )
+
+
 @conversation_blueprint.post("/stream-query")
 @jwt_refresh_token_required
 async def stream_query():
@@ -82,16 +153,31 @@ async def stream_query():
     data = await request.get_json()
     query_text = data.get("query")
     conversation_id = data.get("conversation_id")
+    image_key = data.get("image_key")
     conversation = await get_conversation_by_id(conversation_id)
     await conversation.fetch_related("messages")
     await add_message_to_conversation(
         conversation=conversation,
-        message=query_text,
+        message=query_text or "",
         speaker="user",
         user=user,
+        image_key=image_key,
     )
 
-    messages_payload = _build_messages_payload(conversation, query_text)
+    # If an image was uploaded, analyze it with the vision model
+    image_description = None
+    if image_key:
+        try:
+            image_bytes, _ = await s3_get_image(image_key)
+            image_description = await analyze_user_image(image_bytes)
+            logging.info(f"Image analysis complete for {image_key}")
+        except Exception as e:
+            logging.error(f"Failed to analyze image: {e}")
+            image_description = "[Image could not be analyzed]"
+
+    messages_payload = _build_messages_payload(
+        conversation, query_text or "", image_description
+    )
     payload = {"messages": messages_payload}
 
     async def event_generator():
@@ -160,6 +246,7 @@ async def get_conversation(conversation_id: str):
                 "text": msg.text,
                 "speaker": msg.speaker.value,
                 "created_at": msg.created_at.isoformat(),
+                "image_key": msg.image_key,
             }
         )
     name = conversation.name

blueprints/conversation/logic.py

@@ -16,12 +16,14 @@ async def add_message_to_conversation(
     message: str,
     speaker: str,
     user: blueprints.users.models.User,
+    image_key: str | None = None,
 ) -> ConversationMessage:
     print(conversation, message, speaker)
     message = await ConversationMessage.create(
         text=message,
         speaker=speaker,
         conversation=conversation,
+        image_key=image_key,
     )
 
     return message

blueprints/conversation/models.py

@@ -41,6 +41,7 @@ class ConversationMessage(Model):
     )
     created_at = fields.DatetimeField(auto_now_add=True)
     speaker = fields.CharEnumField(enum_type=Speaker, max_length=10)
+    image_key = fields.CharField(max_length=512, null=True, default=None)
 
     class Meta:
         table = "conversation_messages"

config/db.py

@@ -1,15 +1,14 @@
 import os
+
 from dotenv import load_dotenv
 
-# Load environment variables
 load_dotenv()
 
-# Database configuration with environment variable support
-# Use DATABASE_PATH for relative paths or DATABASE_URL for full connection strings
-DATABASE_PATH = os.getenv("DATABASE_PATH", "database/raggr.db")
-DATABASE_URL = os.getenv("DATABASE_URL", f"sqlite://{DATABASE_PATH}")
+DATABASE_URL = os.getenv(
+    "DATABASE_URL", "postgres://raggr:raggr_dev_password@localhost:5432/raggr"
+)
 
-TORTOISE_ORM = {
+TORTOISE_CONFIG = {
     "connections": {"default": DATABASE_URL},
     "apps": {
         "models": {

docker-compose.yml

@@ -55,6 +55,12 @@ services:
       - OBSIDIAN_DEVICE_NAME=${OBSIDIAN_DEVICE_NAME}
       - OBSIDIAN_CONTINUOUS_SYNC=${OBSIDIAN_CONTINUOUS_SYNC:-false}
       - OBSIDIAN_VAULT_PATH=${OBSIDIAN_VAULT_PATH:-/app/data/obsidian}
+      - S3_ENDPOINT_URL=${S3_ENDPOINT_URL}
+      - S3_ACCESS_KEY_ID=${S3_ACCESS_KEY_ID}
+      - S3_SECRET_ACCESS_KEY=${S3_SECRET_ACCESS_KEY}
+      - S3_BUCKET_NAME=${S3_BUCKET_NAME:-asksimba-images}
+      - S3_REGION=${S3_REGION:-garage}
+      - OLLAMA_HOST=${OLLAMA_HOST:-http://localhost:11434}
     depends_on:
       postgres:
         condition: service_healthy

migrations/models/3_20260313000000_add_email_fields.py

@@ -15,3 +15,32 @@ async def downgrade(db: BaseDBAsyncClient) -> str:
         DROP INDEX IF EXISTS "idx_users_email_h_a1b2c3";
         ALTER TABLE "users" DROP COLUMN "email_hmac_token";
         ALTER TABLE "users" DROP COLUMN "email_enabled";"""
+
+
+MODELS_STATE = (
+    "eJztmm1v4jgQx78Kyquu1KtaKN1VdTopUHrLbYEThX3qVZFJXMg1sbOJsxRV/e5nm4Q4jg"
+    "OEAoU93rRl7CH2z2PP35M+ay62oBOc1DH6Cf0AEBsj7bL0rCHgQvqHsv24pAHPS1qZgYCB"
+    "wx1MoSdvAYOA+MAktPEBOAGkJgsGpm970cNQ6DjMiE3a0UbDxBQi+0cIDYKHkIygTxvu7q"
+    "nZRhZ8gkH80Xs0HmzoWKlx2xZ7NrcbZOJxW7/fvLrmPdnjBoaJndBFSW9vQkYYzbqHoW2d"
+    "MB/WNoQI+oBAS5gGG2U07dg0HTE1ED+Es6FaicGCDyB0GAzt94cQmYxBiT+J/Tj/QyuAh6"
+    "JmaG1EGIvnl+mskjlzq8YeVf+od48qF+/4LHFAhj5v5ES0F+4ICJi6cq4JSP47g7I+Ar4a"
+    "ZdxfgkkHugrG2JBwTGIoBhkDWo2a5oInw4FoSEb0Y7lanYPxs97lJGkvjhLTuJ5GfTtqKk"
+    "/bGNIEoelDNmUDkCzIK9pCbBeqYaY9JaRW5HoS/7GjgOkcrA5yJtEmmMO312w1bnt66282"
+    "EzcIfjgckd5rsJYyt04k69GFtBSzLyl9afY+ltjH0vdOuyHH/qxf77vGxgRCgg2ExwawhP"
+    "0aW2MwqYUNPWvFhU17Hhb2TRc2GrywrgH0jWIZRHB5RRqJxrbFRVw9abDU+/CozBkMRhbe"
+    "NfahPUSf4IQjbNJxAGSqkkUkOvrR1+wqtMSajMIH45kaEYOCzo7OCZJp9tRv6/pVQ+MMB8"
+    "B8HAPfMnJgujAIwBAGWaC1yPP6Uxc6M2mmZikKuNb0G3fzVMljy1nhMhYYpehlm9yyK1sA"
+    "ovO2omezJ82hs0AFCxCXE8OGuJAHUbzXopjAJ0XK71GrGmXcf19E8bxU3vjaS2XxWPoetf"
+    "Sv71KZ/KbT/jPuLkjl+k2ndlDIv6KQyirkwIPgUSUG2AWygUI3IwVSqyu4v/HW0fq3je5l"
+    "iWX0f9Bts1XTL0uB7Q6AttwSp26ZZ6dLXDLPTnPvmKxJ2kBioil2zCtc13nm76mENaWC1y"
+    "ulrFw/21mKCzWtIlyKattNKjl+Z1BIt/guka/V2NY+aLP912ZsHYsWLUWffdFoWyhceiAI"
+    "xthXRGbNRsCfqGGKXhLMwYRM7z+7eqVXwasxvSrKLYqs1mzr3W9qyRv3F+O29q3X0CW60A"
+    "W2UyRKZw7rCdHFO36dAXp2upzomad6MrJnPAIkoEe6QZXkIE9mqmEqXFfCKofqdqlWloFa"
+    "yWdaySDlQWZAxKan2vgYOxCgOQEq+srbnzpv6jAtmqoL7P9O5ya1/2tN+Urbb9UaNHg5Zt"
+    "rJnkqhZrunhDtygUk1wiNUKMsFu1/y3cOIPbtY5hiQr6zCKXAhRyy2LdMIwsG/0FSUD/KB"
+    "yn57CHMjWZ9e6EeG5+OftlXsSM04bk9KaQ42gfMKLZrmWl3mWK3mH6vVzLHqWMAzhj4OPU"
+    "Uh/6/bTluNVHKTgPYRneWdZZvkuOTYAbnfGN67+83ofDbz+dVEuXAoCSv2BYdq4v+kmnh4"
+    "3/5LLOzsdV6mKrToXWjmn8vW80J0l2+k230RqkPfNkeaooAWtRzPK6GBpM/O1NCaKOednL"
+    "KExjBLwRCt/JvepPnr6N/KZ+fvzz9ULs4/0C58JDPL+zmHQXwNyS+ZsY2grHPnaz3B5VAw"
+    "S6Qz3RpFBPO0+34C3EhBhz6RQKRI7/kSWXB5K3m8sdLj2uRxgWy7/vTy8h9Mf/k3"
+)

migrations/models/4_20260404080201_add_image_key.py

@@ -0,0 +1,43 @@
+from tortoise import BaseDBAsyncClient
+
+RUN_IN_TRANSACTION = True
+
+
+async def upgrade(db: BaseDBAsyncClient) -> str:
+    return """
+        ALTER TABLE "conversation_messages" ADD "image_key" VARCHAR(512);"""
+
+
+async def downgrade(db: BaseDBAsyncClient) -> str:
+    return """
+        ALTER TABLE "conversation_messages" DROP COLUMN "image_key";"""
+
+
+MODELS_STATE = (
+    "eJztmmtv4jgUhv8KyqeO1K0KvcyoWq0UWrrDToFVC3PrVpFJXPCS2JnEGYqq/ve1TUIcx6"
+    "GkBQqzfGnLsQ+xH1/Oe076aHjEgW54cE7wTxiEgCKCjbPKo4GBB9kf2vb9igF8P23lBgr6"
+    "rnCwpZ6iBfRDGgCbssZ74IaQmRwY2gHy44fhyHW5kdisI8KD1BRh9COCFiUDSIcwYA23d8"
+    "yMsAMfYJh89EfWPYKukxk3cvizhd2iE1/Yer3mxaXoyR/Xt2ziRh5Oe/sTOiR41j2KkHPA"
+    "fXjbAGIYAAodaRp8lPG0E9N0xMxAgwjOhuqkBgfeg8jlMIzf7yNscwYV8ST+4/gPowQehp"
+    "qjRZhyFo9P01mlcxZWgz/q/KN5vXd0+k7MkoR0EIhGQcR4Eo6Agqmr4JqCFL9zKM+HINCj"
+    "TPorMNlAX4IxMaQc0z2UgEwAvYya4YEHy4V4QIfsY+3kZA7Gz+a1IMl6CZSE7evprm/HTb"
+    "VpG0eaIrQDyKdsAZoHecFaKPKgHmbWU0HqxK4HyR8bCpjNwelgdxIfgjl8u81W46Zrtv7m"
+    "M/HC8IcrEJndBm+pCetEse6dKksx+5LKl2b3Y4V/rHzvtBvq3p/16343+JhARImFydgCjn"
+    "ReE2sCJrOwke+8cGGznruFfdOFjQcvrWsIA6tcBJFcXhFG4rGtcRFfHjR46L0faWMGh5GH"
+    "d0kCiAb4E5wIhE02DoBtXbCIRUcv/ppNhZZa01EEYDxTI/KmYLNjc4J0Gj3Nm3PzomEIhn"
+    "1gj8YgcKwCmB4MQzCAYR5oPfa8/HQN3Zk007OUBVxr+o2beasUsRWsSI1IjDL08k1ezVMt"
+    "ALN5O/Gz+ZPm0HlGBUsQFxPDlryQO1G81aKYwgdNyO8yqx5l0n9bRPG8UN742s1E8UT67r"
+    "XMr+8ykfyq0/4z6S5J5fOrTn2nkH9FIZVXyKEPwUgnBngC2cCRl5MCmdWV3N/46Bi9m8b1"
+    "WYVH9H/wTbNVN88qIfL6wFhsiTNZZvVwgSSzeliYY/Km7AFCHoss1ghOyqTqGacX8V2/9M"
+    "qCPKnWFiDJehWiFG3KZSQH7XIhU+O6zPi5pemArRQPX5kWqLXIjaX4bH6g2S5l84RVqmKR"
+    "f2lkcJKXFetefk3udO7261y+jmULwLLPtujdNRSBfRCGYxJodmYdYRBM9DBlLwVmf0Knue"
+    "TGxeg58Opc+8vSlSGrN9vm9Td9+pD0l/dt/Vu3YSp0oQeQW2aXzhyWs0WfP/HL3KDVw8UE"
+    "5DwFmZOQ4yGgIbvSLabK+0WSXQ9T47oUObleqkeLQD0qZnqUQyo2mQUxn57u4BPiQoDnbF"
+    "DZVz3+zHlVl2nZUF3i/Hc6V5nzX2+q5YFeq95gm1dgZp3QVAo1210t3KEHbKYRRlCjLJ85"
+    "/YrvFu7Y6uki14Ca/ku3wKm6YwlybCuM+v9CW1OKKQaq+m0hzJVEfRDRoeUH5Cdyyl2pOc"
+    "f1SSnDJTZwX6FFlRx9kWv1pPhaPcldq64DfGsQkMjXvBT566bT1iNV3BSgPcxmeesgm+5X"
+    "XBTSu5Xhvb1bjc7nM59fmVWLsIqw4l+wq8z+Tyqzu/9d+CUWdvZqNFcVeu69cu4f9Zbzcn"
+    "mTM9L1vlQ2YYDsoaEpoMUt+/NKaCDtszE1tCYueL+pLaFxzMpmiFf+TTNp8Wr/t1r1+P3x"
+    "h6PT4w+sixjJzPJ+zmWQpCHFJTN+ELR17mKtJ7nsCmapdGZHo4xgnnbfToArKeiwJ1KINe"
+    "G9WCJLLm8lj1dWelyaPC4RbZcfXp7+AzcBYwM="
+)

pyproject.toml

@@ -37,9 +37,10 @@ dependencies = [
     "ynab>=1.3.0",
     "ollama>=0.6.1",
     "twilio>=9.10.2",
+    "aioboto3>=13.0.0",
 ]
 
 [tool.aerich]
-tortoise_orm = "app.TORTOISE_CONFIG"
+tortoise_orm = "config.db.TORTOISE_CONFIG"
 location = "./migrations"
 src_folder = "./."

raggr-frontend/src/api/conversationService.ts

@@ -13,6 +13,7 @@ interface Message {
   text: string;
   speaker: "user" | "simba";
   created_at: string;
+  image_key?: string | null;
 }
 
 interface Conversation {
@@ -121,17 +122,52 @@ class ConversationService {
     return await response.json();
   }
 
+  async uploadImage(
+    file: File,
+    conversationId: string,
+  ): Promise<{ image_key: string; image_url: string }> {
+    const formData = new FormData();
+    formData.append("file", file);
+    formData.append("conversation_id", conversationId);
+
+    const response = await userService.fetchWithRefreshToken(
+      `${this.conversationBaseUrl}/upload-image`,
+      {
+        method: "POST",
+        body: formData,
+      },
+      { skipContentType: true },
+    );
+
+    if (!response.ok) {
+      const data = await response.json();
+      throw new Error(data.error || "Failed to upload image");
+    }
+
+    return await response.json();
+  }
+
+  getImageUrl(imageKey: string): string {
+    return `/api/conversation/image/${imageKey}`;
+  }
+
   async streamQuery(
     query: string,
     conversation_id: string,
     onEvent: SSEEventCallback,
     signal?: AbortSignal,
+    imageKey?: string,
   ): Promise<void> {
+    const body: Record<string, string> = { query, conversation_id };
+    if (imageKey) {
+      body.image_key = imageKey;
+    }
+
     const response = await userService.fetchWithRefreshToken(
       `${this.conversationBaseUrl}/stream-query`,
       {
         method: "POST",
-        body: JSON.stringify({ query, conversation_id }),
+        body: JSON.stringify(body),
         signal,
       },
     );

raggr-frontend/src/api/userService.ts

@@ -106,14 +106,15 @@ class UserService {
   async fetchWithRefreshToken(
     url: string,
     options: RequestInit = {},
+    { skipContentType = false }: { skipContentType?: boolean } = {},
   ): Promise<Response> {
     const refreshToken = localStorage.getItem("refresh_token");
 
     // Add authorization header
-    const headers = {
-      "Content-Type": "application/json",
-      ...(options.headers || {}),
-      ...(refreshToken && { Authorization: `Bearer ${refreshToken}` }),
+    const headers: Record<string, string> = {
+      ...(skipContentType ? {} : { "Content-Type": "application/json" }),
+      ...((options.headers as Record<string, string>) || {}),
+      ...(refreshToken ? { Authorization: `Bearer ${refreshToken}` } : {}),
     };
 
     let response = await fetch(url, { ...options, headers });

raggr-frontend/src/components/ChatScreen.tsx

@@ -14,6 +14,7 @@ import catIcon from "../assets/cat.png";
 type Message = {
   text: string;
   speaker: "simba" | "user" | "tool";
+  image_key?: string | null;
 };
 
 type Conversation = {
@@ -55,6 +56,7 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
   const [isLoading, setIsLoading] = useState<boolean>(false);
   const [isAdmin, setIsAdmin] = useState<boolean>(false);
   const [showAdminPanel, setShowAdminPanel] = useState<boolean>(false);
+  const [pendingImage, setPendingImage] = useState<File | null>(null);
 
   const messagesEndRef = useRef<HTMLDivElement>(null);
   const isMountedRef = useRef<boolean>(true);
@@ -80,7 +82,7 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
       try {
         const fetched = await conversationService.getConversation(conversation.id);
         setMessages(
-          fetched.messages.map((m) => ({ text: m.text, speaker: m.speaker })),
+          fetched.messages.map((m) => ({ text: m.text, speaker: m.speaker, image_key: m.image_key })),
         );
       } catch (err) {
         console.error("Failed to load messages:", err);
@@ -120,7 +122,7 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
       try {
         const conv = await conversationService.getConversation(selectedConversation.id);
         setSelectedConversation({ id: conv.id, title: conv.name });
-        setMessages(conv.messages.map((m) => ({ text: m.text, speaker: m.speaker })));
+        setMessages(conv.messages.map((m) => ({ text: m.text, speaker: m.speaker, image_key: m.image_key })));
       } catch (err) {
         console.error("Failed to load messages:", err);
       }
@@ -129,7 +131,7 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
   }, [selectedConversation?.id]);
 
   const handleQuestionSubmit = async () => {
-    if (!query.trim() || isLoading) return;
+    if ((!query.trim() && !pendingImage) || isLoading) return;
 
     let activeConversation = selectedConversation;
     if (!activeConversation) {
@@ -139,9 +141,13 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
       setConversations((prev) => [activeConversation!, ...prev]);
     }
 
+    // Capture pending image before clearing state
+    const imageFile = pendingImage;
+
     const currMessages = messages.concat([{ text: query, speaker: "user" }]);
     setMessages(currMessages);
     setQuery("");
+    setPendingImage(null);
     setIsLoading(true);
 
     if (simbaMode) {
@@ -155,6 +161,29 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
     abortControllerRef.current = abortController;
 
     try {
+      // Upload image first if present
+      let imageKey: string | undefined;
+      if (imageFile) {
+        const uploadResult = await conversationService.uploadImage(
+          imageFile,
+          activeConversation.id,
+        );
+        imageKey = uploadResult.image_key;
+
+        // Update the user message with the image key
+        setMessages((prev) => {
+          const updated = [...prev];
+          // Find the last user message we just added
+          for (let i = updated.length - 1; i >= 0; i--) {
+            if (updated[i].speaker === "user") {
+              updated[i] = { ...updated[i], image_key: imageKey };
+              break;
+            }
+          }
+          return updated;
+        });
+      }
+
       await conversationService.streamQuery(
         query,
         activeConversation.id,
@@ -170,6 +199,7 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
           }
         },
         abortController.signal,
+        imageKey,
       );
     } catch (error) {
       if (error instanceof Error && error.name === "AbortError") {
@@ -349,6 +379,9 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
                 handleQuestionSubmit={handleQuestionSubmit}
                 setSimbaMode={setSimbaMode}
                 isLoading={isLoading}
+                pendingImage={pendingImage}
+                onImageSelect={(file) => setPendingImage(file)}
+                onClearImage={() => setPendingImage(null)}
               />
             </div>
           </div>
@@ -375,7 +408,7 @@ export const ChatScreen = ({ setAuthenticated }: ChatScreenProps) => {
                     return <ToolBubble key={index} text={msg.text} />;
                   if (msg.speaker === "simba")
                     return <AnswerBubble key={index} text={msg.text} />;
-                  return <QuestionBubble key={index} text={msg.text} />;
+                  return <QuestionBubble key={index} text={msg.text} image_key={msg.image_key} />;
                 })}
 
                 {isLoading && <AnswerBubble text="" loading={true} />}

raggr-frontend/src/components/MessageInput.tsx

@@ -1,5 +1,5 @@
-import { useState } from "react";
-import { ArrowUp } from "lucide-react";
+import { useRef, useState } from "react";
+import { ArrowUp, ImagePlus, X } from "lucide-react";
 import { cn } from "../lib/utils";
 import { Textarea } from "./ui/textarea";
 
@@ -10,6 +10,9 @@ type MessageInputProps = {
   setSimbaMode: (val: boolean) => void;
   query: string;
   isLoading: boolean;
+  pendingImage: File | null;
+  onImageSelect: (file: File) => void;
+  onClearImage: () => void;
 };
 
 export const MessageInput = ({
@@ -19,8 +22,12 @@ export const MessageInput = ({
   handleQuestionSubmit,
   setSimbaMode,
   isLoading,
+  pendingImage,
+  onImageSelect,
+  onClearImage,
 }: MessageInputProps) => {
   const [simbaMode, setLocalSimbaMode] = useState(false);
+  const fileInputRef = useRef<HTMLInputElement>(null);
 
   const toggleSimbaMode = () => {
     const next = !simbaMode;
@@ -28,6 +35,17 @@ export const MessageInput = ({
     setSimbaMode(next);
   };
 
+  const handleFileChange = (e: React.ChangeEvent<HTMLInputElement>) => {
+    const file = e.target.files?.[0];
+    if (file) {
+      onImageSelect(file);
+    }
+    // Reset so the same file can be re-selected
+    e.target.value = "";
+  };
+
+  const canSend = !isLoading && (query.trim() || pendingImage);
+
   return (
     <div
       className={cn(
@@ -36,6 +54,26 @@ export const MessageInput = ({
         "focus-within:border-amber-soft/60",
       )}
     >
+      {/* Image preview */}
+      {pendingImage && (
+        <div className="px-3 pt-3">
+          <div className="relative inline-block">
+            <img
+              src={URL.createObjectURL(pendingImage)}
+              alt="Pending upload"
+              className="h-20 rounded-lg object-cover border border-sand"
+            />
+            <button
+              type="button"
+              onClick={onClearImage}
+              className="absolute -top-1.5 -right-1.5 w-5 h-5 rounded-full bg-charcoal text-white flex items-center justify-center hover:bg-charcoal/80 transition-colors cursor-pointer"
+            >
+              <X size={12} />
+            </button>
+          </div>
+        </div>
+      )}
+
       {/* Textarea */}
       <Textarea
         onChange={handleQueryChange}
@@ -46,8 +84,18 @@ export const MessageInput = ({
         className="min-h-[60px] max-h-40"
       />
 
+      {/* Hidden file input */}
+      <input
+        ref={fileInputRef}
+        type="file"
+        accept="image/*"
+        onChange={handleFileChange}
+        className="hidden"
+      />
+
       {/* Bottom toolbar */}
       <div className="flex items-center justify-between px-3 pb-2.5 pt-1">
+        <div className="flex items-center gap-3">
           {/* Simba mode toggle */}
           <button
             type="button"
@@ -62,16 +110,32 @@ export const MessageInput = ({
             </span>
           </button>
 
+          {/* Image attach button */}
+          <button
+            type="button"
+            onClick={() => fileInputRef.current?.click()}
+            disabled={isLoading}
+            className={cn(
+              "w-7 h-7 rounded-lg flex items-center justify-center transition-all cursor-pointer",
+              isLoading
+                ? "text-warm-gray/40 cursor-not-allowed"
+                : "text-warm-gray hover:text-charcoal hover:bg-cream-dark",
+            )}
+          >
+            <ImagePlus size={16} />
+          </button>
+        </div>
+
         {/* Send button */}
         <button
           type="submit"
           onClick={handleQuestionSubmit}
-          disabled={isLoading || !query.trim()}
+          disabled={!canSend}
           className={cn(
             "w-8 h-8 rounded-full flex items-center justify-center",
             "transition-all duration-200 cursor-pointer",
             "shadow-sm",
-            isLoading || !query.trim()
+            !canSend
               ? "bg-sand text-warm-gray/50 cursor-not-allowed shadow-none"
               : "bg-amber-glow text-white hover:bg-amber-dark hover:shadow-md hover:shadow-amber-glow/30 active:scale-95",
           )}

raggr-frontend/src/components/QuestionBubble.tsx

@@ -1,10 +1,12 @@
 import { cn } from "../lib/utils";
+import { conversationService } from "../api/conversationService";
 
 type QuestionBubbleProps = {
   text: string;
+  image_key?: string | null;
 };
 
-export const QuestionBubble = ({ text }: QuestionBubbleProps) => {
+export const QuestionBubble = ({ text, image_key }: QuestionBubbleProps) => {
   return (
     <div className="flex justify-end message-enter">
       <div
@@ -15,6 +17,13 @@ export const QuestionBubble = ({ text }: QuestionBubbleProps) => {
           "shadow-sm shadow-leaf/10",
         )}
       >
+        {image_key && (
+          <img
+            src={conversationService.getImageUrl(image_key)}
+            alt="Uploaded image"
+            className="max-w-full rounded-xl mb-2"
+          />
+        )}
         {text}
       </div>
     </div>

utils/image_process.py

@@ -76,6 +76,50 @@ def describe_simba_image(input):
     return result
 
 
+async def analyze_user_image(file_bytes: bytes) -> str:
+    """Analyze an image uploaded by a user and return a text description.
+
+    Uses llama-server (OpenAI-compatible API) with vision support.
+    Falls back to OpenAI if llama-server is not configured.
+    """
+    import base64
+
+    from openai import AsyncOpenAI
+
+    llama_url = os.getenv("LLAMA_SERVER_URL")
+    if llama_url:
+        aclient = AsyncOpenAI(base_url=llama_url, api_key="not-needed")
+        model = os.getenv("LLAMA_MODEL_NAME", "llama-3.1-8b-instruct")
+    else:
+        aclient = AsyncOpenAI()
+        model = "gpt-4o-mini"
+
+    b64 = base64.b64encode(file_bytes).decode("utf-8")
+
+    response = await aclient.chat.completions.create(
+        model=model,
+        messages=[
+            {
+                "role": "system",
+                "content": "You are a helpful image analyst. Describe what you see in the image in detail. Be thorough but concise.",
+            },
+            {
+                "role": "user",
+                "content": [
+                    {"type": "text", "text": "Please describe this image in detail."},
+                    {
+                        "type": "image_url",
+                        "image_url": {
+                            "url": f"data:image/jpeg;base64,{b64}",
+                        },
+                    },
+                ],
+            },
+        ],
+    )
+    return response.choices[0].message.content
+
+
 if __name__ == "__main__":
     args = parser.parse_args()
     if args.filepath:

utils/image_upload.py

@@ -0,0 +1,62 @@
+import io
+import logging
+
+from PIL import Image
+from pillow_heif import register_heif_opener
+
+register_heif_opener()
+
+logging.basicConfig(level=logging.INFO)
+
+ALLOWED_TYPES = {"image/jpeg", "image/png", "image/webp", "image/heic", "image/heif"}
+MAX_DIMENSION = 1920
+
+
+class ImageValidationError(Exception):
+    pass
+
+
+def process_image(file_bytes: bytes, content_type: str) -> tuple[bytes, str]:
+    """Validate, resize, and strip EXIF from an uploaded image.
+
+    Returns processed bytes and the output content type (always image/jpeg or image/png or image/webp).
+    """
+    if content_type not in ALLOWED_TYPES:
+        raise ImageValidationError(
+            f"Unsupported image type: {content_type}. "
+            f"Allowed: JPEG, PNG, WebP, HEIC"
+        )
+
+    img = Image.open(io.BytesIO(file_bytes))
+
+    # Resize if too large
+    width, height = img.size
+    if max(width, height) > MAX_DIMENSION:
+        ratio = MAX_DIMENSION / max(width, height)
+        new_size = (int(width * ratio), int(height * ratio))
+        img = img.resize(new_size, Image.LANCZOS)
+        logging.info(
+            f"Resized image from {width}x{height} to {new_size[0]}x{new_size[1]}"
+        )
+
+    # Strip EXIF by copying pixel data to a new image
+    clean_img = Image.new(img.mode, img.size)
+    clean_img.putdata(list(img.getdata()))
+
+    # Convert HEIC/HEIF to JPEG; otherwise keep original format
+    if content_type in {"image/heic", "image/heif"}:
+        output_format = "JPEG"
+        output_content_type = "image/jpeg"
+    elif content_type == "image/png":
+        output_format = "PNG"
+        output_content_type = "image/png"
+    elif content_type == "image/webp":
+        output_format = "WEBP"
+        output_content_type = "image/webp"
+    else:
+        output_format = "JPEG"
+        output_content_type = "image/jpeg"
+
+    buf = io.BytesIO()
+    clean_img.save(buf, format=output_format, quality=85)
+    return buf.getvalue(), output_content_type

utils/s3_client.py

@@ -0,0 +1,53 @@
+import os
+import logging
+
+import aioboto3
+from dotenv import load_dotenv
+
+load_dotenv()
+
+logging.basicConfig(level=logging.INFO)
+
+S3_ENDPOINT_URL = os.getenv("S3_ENDPOINT_URL")
+S3_ACCESS_KEY_ID = os.getenv("S3_ACCESS_KEY_ID")
+S3_SECRET_ACCESS_KEY = os.getenv("S3_SECRET_ACCESS_KEY")
+S3_BUCKET_NAME = os.getenv("S3_BUCKET_NAME", "asksimba-images")
+S3_REGION = os.getenv("S3_REGION", "garage")
+
+session = aioboto3.Session()
+
+
+def _get_client():
+    return session.client(
+        "s3",
+        endpoint_url=S3_ENDPOINT_URL,
+        aws_access_key_id=S3_ACCESS_KEY_ID,
+        aws_secret_access_key=S3_SECRET_ACCESS_KEY,
+        region_name=S3_REGION,
+    )
+
+
+async def upload_image(file_bytes: bytes, key: str, content_type: str) -> str:
+    async with _get_client() as client:
+        await client.put_object(
+            Bucket=S3_BUCKET_NAME,
+            Key=key,
+            Body=file_bytes,
+            ContentType=content_type,
+        )
+    logging.info(f"Uploaded image to S3: {key}")
+    return key
+
+
+async def get_image(key: str) -> tuple[bytes, str]:
+    async with _get_client() as client:
+        response = await client.get_object(Bucket=S3_BUCKET_NAME, Key=key)
+        body = await response["Body"].read()
+        content_type = response.get("ContentType", "image/jpeg")
+    return body, content_type
+
+
+async def delete_image(key: str) -> None:
+    async with _get_client() as client:
+        await client.delete_object(Bucket=S3_BUCKET_NAME, Key=key)
+    logging.info(f"Deleted image from S3: {key}")

utils/ynab_service.py

@@ -82,7 +82,6 @@ class YNABService:
         end_date: Optional[str] = None,
         category_name: Optional[str] = None,
         payee_name: Optional[str] = None,
-        limit: int = 50,
     ) -> dict[str, Any]:
         """Get transactions filtered by date range, category, or payee.
 
@@ -91,7 +90,6 @@ class YNABService:
             end_date: End date in YYYY-MM-DD format (defaults to today)
             category_name: Filter by category name (case-insensitive partial match)
             payee_name: Filter by payee name (case-insensitive partial match)
-            limit: Maximum number of transactions to return (default 50)
 
         Returns:
             Dictionary containing matching transactions and summary statistics.
@@ -145,9 +143,8 @@ class YNABService:
             )
             total_amount += amount
 
-        # Sort by date (most recent first) and limit
+        # Sort by date (most recent first)
         filtered_transactions.sort(key=lambda x: x["date"], reverse=True)
-        filtered_transactions = filtered_transactions[:limit]
 
         return {
             "transactions": filtered_transactions,