diff --git a/poetry.lock b/poetry.lock
index afedf77..fc2d752 100644
--- a/poetry.lock
+++ b/poetry.lock
@@ -3756,6 +3756,29 @@ pydantic = "<3.0"
 pylint = ">2.0,<3.0"
 pylint-plugin-utils = "*"
 
+[[package]]
+name = "pypdf"
+version = "5.4.0"
+description = "A pure-python PDF library capable of splitting, merging, cropping, and transforming PDF files"
+optional = false
+python-versions = ">=3.8"
+groups = ["main"]
+files = [
+    {file = "pypdf-5.4.0-py3-none-any.whl", hash = "sha256:db994ab47cadc81057ea1591b90e5b543e2b7ef2d0e31ef41a9bfe763c119dab"},
+    {file = "pypdf-5.4.0.tar.gz", hash = "sha256:9af476a9dc30fcb137659b0dec747ea94aa954933c52cf02ee33e39a16fe9175"},
+]
+
+[package.dependencies]
+typing_extensions = {version = ">=4.0", markers = "python_version < \"3.11\""}
+
+[package.extras]
+crypto = ["cryptography"]
+cryptodome = ["PyCryptodome"]
+dev = ["black", "flit", "pip-tools", "pre-commit (<2.18.0)", "pytest-cov", "pytest-socket", "pytest-timeout", "pytest-xdist", "wheel"]
+docs = ["myst_parser", "sphinx", "sphinx_rtd_theme"]
+full = ["Pillow (>=8.0.0)", "cryptography"]
+image = ["Pillow (>=8.0.0)"]
+
 [[package]]
 name = "pysbd"
 version = "0.3.4"
@@ -5959,4 +5982,4 @@ testing = ["coverage[toml]", "zope.event", "zope.testing"]
 [metadata]
 lock-version = "2.1"
 python-versions = ">=3.10,<3.13"
-content-hash = "395a5f09607b7ca5550b4f6f6002fdd8104f1e374f41e83701f8991c59198242"
+content-hash = "9368cb6daecdfaa9dd3d28cd0234989e076b063c69711804b892dd6149a83afa"
diff --git a/pyproject.toml b/pyproject.toml
index 038f711..bb67594 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -35,6 +35,7 @@ uvicorn = {extras = ["standard"], version = "^0.29.0"}
 autogen-core = "^0.4.9.3"
 autogen-ext = "^0.4.9.3"
 autogen-agentchat = "^0.4.9.2"
+pypdf = "^5.4.0"
 
 [tool.poetry.group.dev.dependencies]
 pytest = "^7.4.0"
diff --git a/src/app/api/api_v1/endpoints/tutor.py b/src/app/api/api_v1/endpoints/tutor.py
index 293b862..9cc9601 100644
--- a/src/app/api/api_v1/endpoints/tutor.py
+++ b/src/app/api/api_v1/endpoints/tutor.py
@@ -1,6 +1,7 @@
 from typing import Annotated
 
 from fastapi import APIRouter, File, Response, UploadFile
+from pypdf import PdfReader
 
 from src.app.api.dependencies import get_settings
 from src.app.services.abst_chat import AbstractChat, ChatFactory
@@ -20,7 +21,7 @@
 
 settings = get_settings()
 
-chatfactory: AbstractChat = ChatFactory().create_chat("openai")
+chatfactory: AbstractChat = ChatFactory().create_chat("openai", model="gpt-4o")
 chatfactory.init_client()
 
 sp = SearchService()
@@ -39,7 +40,21 @@ async def tutor_search(
     files: Annotated[list[UploadFile], File()],
     response: Response,
 ):
-    file_content: list[bytes] = [await file.read() for file in files]
+    files_content: list[bytes] = []
+    for file in files:
+        if (
+            file.content_type == "application/pdf"
+            or file.content_type == "application/x-pdf"
+        ):
+            file_content = ""
+            reader = PdfReader(file.file)
+            for page in reader.pages:
+                file_content += page.extract_text()
+            files_content.append(file_content.encode("utf-8", errors="ignore"))
+        else:
+            file_content = await file.read()
+            files_content.append(file_content)
+
     doc_list_to_string = "Document {doc_nb}: {content}"
 
     file_content_str = [
@@ -47,18 +62,29 @@ async def tutor_search(
             doc_nb=index + 1,
             content=content.decode("utf-8", errors="ignore"),
         )
-        for index, content in enumerate(file_content)
+        for index, content in enumerate(files_content)
     ]
     file_content_str = "\n\n".join(file_content_str)
 
+    print(file_content_str)
+
     messages = [
         {"role": "system", "content": extractor_prompt},
         {"role": "assistant", "content": file_content_str},
     ]
 
-    themes_extracted = await chatfactory.chat_schema(
-        model="gpt-4o-mini", messages=messages, response_format=ExtractorOuputList  # type: ignore
-    )
+    try:
+        themes_extracted = await chatfactory.chat_schema(
+            model="gpt-4o", messages=messages, response_format=ExtractorOuputList  # type: ignore
+        )
+    except Exception as e:
+        logger.error(f"Error in chat schema: {e}")
+        # todo: handle error
+        return TutorSearchResponse(
+            extracts=[],
+            nb_results=0,
+            documents=[],
+        )
 
     if not themes_extracted or not themes_extracted.extracts:
         return TutorSearchResponse(
diff --git a/src/app/services/abst_chat.py b/src/app/services/abst_chat.py
index f1a9978..f967c9a 100644
--- a/src/app/services/abst_chat.py
+++ b/src/app/services/abst_chat.py
@@ -19,12 +19,12 @@
 from abc import ABC, abstractmethod
 from typing import AsyncIterable, Dict, List, Literal, Optional
 
-import openai
 from azure.ai.inference import ChatCompletionsClient
 from azure.core.credentials import AzureKeyCredential
 
 # from ecologits import EcoLogits  # type: ignore
 from mistralai import Mistral
+from openai import AsyncAzureOpenAI
 from pydantic import BaseModel
 
 from src.app.api.dependencies import get_settings
@@ -456,7 +456,7 @@ def init_client(self):
             raise ValueError("API_BASE or API_VERSION not provided")
 
         try:
-            self.chat_client = openai.AsyncAzureOpenAI(
+            self.chat_client = AsyncAzureOpenAI(
                 api_key=self.API_KEY,
                 azure_endpoint=self.API_BASE,
                 api_version=self.API_VERSION,
@@ -507,7 +507,7 @@ async def chat_schema(
     ):
         try:
             completion = await self.chat_client.beta.chat.completions.parse(
-                model=model,
+                model=self.model or model,
                 messages=messages,
                 temperature=0.2,
                 response_format=response_format,
@@ -653,6 +653,26 @@ def create_chat(
             raise ValueError(f"Unsupported chat type: {chat_type}")
 
         if chat_type == "openai":
+            openai_models = {
+                "gpt-4o-mini": (
+                    settings.AZURE_API_KEY,
+                    settings.AZURE_API_BASE,
+                    settings.AZURE_API_VERSION,
+                ),
+                "gpt-4o": (
+                    settings.AZURE_GPT_4O_API_KEY,
+                    settings.AZURE_GPT_4O_API_BASE,
+                    "2025-01-01-preview",
+                ),
+            }
+            if model:
+                key, base, version = openai_models.get(model, (None, None, None))
+                if not key or not base or not version:
+                    raise ValueError(f"Unsupported model: {model}")
+                return chat_classes[chat_type](
+                    API_KEY=key, API_BASE=base, API_VERSION=version, model=model
+                )
+
             return chat_classes[chat_type](
                 API_VERSION=settings.AZURE_API_VERSION,
                 API_KEY=settings.AZURE_API_KEY,
diff --git a/src/app/services/tutor/tutor.py b/src/app/services/tutor/tutor.py
index d68d255..4070dd1 100644
--- a/src/app/services/tutor/tutor.py
+++ b/src/app/services/tutor/tutor.py
@@ -22,7 +22,12 @@
     sdg_expert_topic_type,
     university_teacher_topic_type,
 )
-from src.app.services.tutor.models import Message, TaskResponse, TutorSearchResponse, MessageWithResources
+from src.app.services.tutor.models import (
+    Message,
+    MessageWithResources,
+    TaskResponse,
+    TutorSearchResponse,
+)
 from src.app.services.tutor.utils import extract_doc_info
 
 settings = get_settings()
@@ -40,7 +45,9 @@
 async def tutor_manager(content: TutorSearchResponse) -> Message:
     queue = asyncio.Queue[TaskResponse]()
 
-    formatted_content = MessageWithResources(content=content.extracts, resources=extract_doc_info(content.documents))
+    formatted_content = MessageWithResources(
+        content=content.extracts, resources=extract_doc_info(content.documents)
+    )
 
     async def collect_result(
         _agent: ClosureContext, message: TaskResponse, ctx: MessageContext