diff --git a/__pycache__/deepseek.cpython-313.pyc b/__pycache__/deepseek.cpython-313.pyc
new file mode 100644
index 0000000..5c8b719
Binary files /dev/null and b/__pycache__/deepseek.cpython-313.pyc differ
diff --git a/__pycache__/doubao.cpython-313.pyc b/__pycache__/doubao.cpython-313.pyc
new file mode 100644
index 0000000..403486d
Binary files /dev/null and b/__pycache__/doubao.cpython-313.pyc differ
diff --git a/deepseek.py b/deepseek.py
new file mode 100644
index 0000000..b440529
--- /dev/null
+++ b/deepseek.py
@@ -0,0 +1,64 @@
+# Please install OpenAI SDK first: `pip3 install openai`
+import os
+import json
+from openai import OpenAI
+from dotenv import load_dotenv
+
+load_dotenv()
+
+
+class DeepSeek:
+    system_prompt = (
+        "You are a professional QA answer assistant.\n"
+        "The user will provide a JSON object containing a question.\n"
+        "Your job:\n"
+        "1. Read the 'title' and the 'chooies'.\n"
+        "2. Determine the correct answer(s).\n"
+        "3. Output ONLY the 'num' of the correct choices.\n"
+        "4. If multiple answers exist, join them with English commas, e.g., 'A,C,D'.\n"
+        "5. If you cannot determine the answer confidently, return an empty string.\n"
+        "6. Do not output explanations or any extra text."
+    )
+
+    def __init__(self) -> None:
+        self.key = os.environ.get("DEEPSEEK_API_KEY")
+        if self.key:
+            raise Exception("找不到 DEEPSEEK_API_KEY 请设置环境变量在运行代码")
+        self.client = OpenAI(
+            api_key=self.key,
+            base_url="https://api.deepseek.com",
+        )
+
+    def chat(self, question_json: dict) -> str:
+        response = self.client.chat.completions.create(
+            model="deepseek-chat",
+            # model="deepseek-reasoner",
+            messages=[
+                {"role": "system", "content": self.system_prompt},
+                {"role": "user", "content": json.dumps(question_json)},
+            ],
+            stream=False,
+        )
+
+        # print(response.choices[0].message.content)
+        return (
+            response.choices[0].message.content
+            if response.choices[0].message.content
+            else ""
+        )
+
+
+if __name__ == "__main__":
+    question = {
+        "id": "25594265",
+        "index": "1",
+        "type": "单选",
+        "title": "下列哪个属于AI视觉识别应用？",
+        "chooies": [
+            {"num": "A", "txt": "音乐合成"},
+            {"num": "B", "txt": "植物识别"},
+            {"num": "C", "txt": "文案创作"},
+            {"num": "D", "txt": "机器翻译"},
+        ],
+    }
+    DeepSeek().chat(question)
diff --git a/doubao.py b/doubao.py
new file mode 100644
index 0000000..375393b
--- /dev/null
+++ b/doubao.py
@@ -0,0 +1,65 @@
+# Please install OpenAI SDK first: `pip3 install openai`
+import os
+import json
+from openai import OpenAI
+from dotenv import load_dotenv
+
+load_dotenv()
+
+
+class DOUBAO:
+    system_prompt = (
+        "You are a professional QA answer assistant.\n"
+        "The user will provide a JSON object containing a question.\n"
+        "Your job:\n"
+        "1. Read the 'title' and the 'chooies'.\n"
+        "2. Determine the correct answer(s).\n"
+        "3. Output ONLY the 'num' of the correct choices.\n"
+        "4. If multiple answers exist, join them with English commas, e.g., 'A,C,D'.\n"
+        "5. If you cannot determine the answer confidently, return an empty string.\n"
+        "6. Do not output explanations or any extra text."
+    )
+
+    def __init__(self) -> None:
+        self.key = os.environ.get("DOUBAO_API_KEY")
+        if self.key:
+            raise Exception("找不到 DOUBAO_API_KEY 请设置环境变量在运行代码")
+        self.client = OpenAI(
+            api_key=self.key,
+            base_url="https://ark.cn-beijing.volces.com/api/v3",
+        )
+
+    def chat(self, question_json: dict) -> str:
+        response = self.client.chat.completions.create(
+            # model="deepseek-chat",
+            model="doubao-seed-code-preview-251028",
+            # model="deepseek-reasoner",
+            messages=[
+                {"role": "system", "content": self.system_prompt},
+                {"role": "user", "content": json.dumps(question_json)},
+            ],
+            stream=False,
+        )
+
+        # print(response.choices[0].message.content)
+        return (
+            response.choices[0].message.content
+            if response.choices[0].message.content
+            else ""
+        )
+
+
+if __name__ == "__main__":
+    question = {
+        "id": "25594265",
+        "index": "1",
+        "type": "单选",
+        "title": "下列哪个属于AI视觉识别应用？",
+        "chooies": [
+            {"num": "A", "txt": "音乐合成"},
+            {"num": "B", "txt": "植物识别"},
+            {"num": "C", "txt": "文案创作"},
+            {"num": "D", "txt": "机器翻译"},
+        ],
+    }
+    DOUBAO().chat(question)
diff --git a/main.py b/main.py
index a6db6b1..af15fd1 100644
--- a/main.py
+++ b/main.py
@@ -26,6 +26,9 @@ from requests.adapters import HTTPAdapter
 from urllib3.util.retry import Retry
 import warnings
 
+from deepseek import DeepSeek
+from doubao import DOUBAO
+
 # 修复 ANTIALIAS 错误 - 添加猴子补丁
 import PIL.Image
 
@@ -79,7 +82,20 @@ def replace_html_entities(text, replace_mode="keep_char"):
 
 
 class CKWK:
-    def __init__(self, un: str, pw: str, host: str) -> None:
+    def __init__(self, un: str, pw: str, host: str, ai_type="doubao") -> None:
+        """
+        un: 用户名
+        pw: 密码
+        host: 网站域名
+        ai_type: ai类别取值为 'doubao' 或者 'deepseek'
+        """
+        if ai_type == "doubao":
+            self.ai = DOUBAO()
+        elif ai_type == "deepseek":
+            self.ai = DeepSeek()
+        else:
+            self.ai = DOUBAO()
+
         self.session = requests.Session()
         # 定义重试策略
         retry_strategy = Retry(
@@ -93,18 +109,20 @@ class CKWK:
         adapter = HTTPAdapter(max_retries=retry_strategy)
         self.session.mount("http://", adapter)
         self.session.mount("https://", adapter)
-        # self.session.verify = False
-        # self.session.proxies.update(
-        #     {
-        #         "http": "http://127.0.0.1:9000",
-        #         "https": "http://127.0.0.1:9000",
-        #     }
-        # )
+        self.session.verify = False
+        self.session.proxies.update(
+            {
+                "http": "http://127.0.0.1:9000",
+                "https": "http://127.0.0.1:9000",
+            }
+        )
         self.username = un
         self.host = host
         self.password = pw
         self.user = dict()
         self.courses = []
+        self.works = []
+        self.qas = []
         self.ocr = ddddocr.DdddOcr()
         self.session.headers.update(
             {
@@ -215,6 +233,82 @@ class CKWK:
         finally:
             return video, state
 
+    def parse_q_a(self, html: str) -> None:
+        tree = etree.HTML(html)
+        qas = tree.xpath('//div[@class="topic-item"]/form')
+        for qa in qas:
+            items = {
+                "id": qa.xpath(
+                    './div[@class="courseexamcon-submit"]/input[@name="answerId"]/@value'
+                )[0],
+                "index": qa.xpath(
+                    './div[@class="courseexamcon-main"]/div/div[@class="num"]/span/text()'
+                )[0],
+                "type": qa.xpath(
+                    './div[@class="courseexamcon-main"]/div/div[@class="type"]/text()'
+                )[0],
+                "title": qa.xpath(
+                    './div[@class="courseexamcon-main"]/div[@class="name"]/text()'
+                )[0].strip(),
+                "chooies": [
+                    {
+                        "num": a.xpath('.//span[@class="num"]/text()')[0],
+                        "txt": a.xpath('.//span[@class="txt"]/text()')[0],
+                    }
+                    for a in qa.xpath(
+                        './div[@class="courseexamcon-main"]/div[@class="list"]//li'
+                    )
+                ],
+            }
+            self.qas.append(items)
+
+    def get_question_answer(self, courseID: str, nodeID: str, workID: str):
+        headers = self.session.headers
+        headers["x-requested-with"] = "XMLHttpRequest"
+        resp = self.session.get(
+            "https://{}/user/work/start".format(self.host),
+            params={"courseId": courseID, "nodeId": nodeID, "workId": workID},
+            headers=headers,
+        )
+        if resp.status_code == 200 and resp.json().get("status", False):
+            qa_path = resp.json().get("url", "")
+            self.log(f'开始作业: {resp.json().get("msg", "")} url -> {qa_path}')
+            resp2 = self.session.get("https://{}{}".format(self.host, qa_path))
+            if resp.status_code == 200:
+                self.parse_q_a(resp2.text)
+        else:
+            self.log(f'开始作业: {resp.json().get("msg", "")}')
+            return False
+        return True
+
+    def work_submit(self, answerID: str, workID: str, answer: str, final: bool):
+        headers = self.session.headers
+        headers["x-requested-with"] = "XMLHttpRequest"
+        data = {
+            "answerId": answerID,
+            "workId": workID,
+            "finish": "0" if final else "1",
+        }
+        if "," in answer:
+            answers = answer.split(",")
+            data["answer[]"] = answers
+            # for a in answers:
+            #     data.setdefault("answer[]", []).append(a)
+        else:
+            if answer == "":
+                data["answer"] = "C"
+            data["answer"] = answer
+
+        resp = self.session.post(
+            f"https://{self.host}/user/work/submit",
+            data=data,
+            headers=headers,
+        )
+        if resp.status_code == 200 and resp.json().get("status", False):
+            self.log(
+                f'提交答案: {resp.json().get("msg", "")} {answerID} => {answer} {final}'
+            )
+
     def get_course(self, _id: str) -> tuple[int, int]:
         resp = self.session.get("https://{}/user/node?nodeId={}".format(self.host, _id))
         if resp.status_code == 200 and "错误提示" not in resp.text:
@@ -306,7 +400,23 @@ class CKWK:
         if resp.status_code == 200 and resp.json().get("status", False):
             self.log(f'{resp.json().get("msg", "")} page -> {page}')
             self.courses.extend(resp.json().get("list", []))
-            if page <= resp.json().get("pageInfo", {}).get("pageCount", 0):
+            if page < resp.json().get("pageInfo", {}).get("pageCount", 0):
+                self.get_study_record(course_id, page + 1)
+
+    def get_work_record(self, course_id: str, user_id: str, page=1):
+        resp = self.session.get(
+            f"https://{self.host}/user/study_record/work.json",
+            params={
+                "courseId": course_id,
+                # "userId": user_id,
+                "page": page,
+                "_": str(int(time.time() * 1000)),
+            },
+        )
+        if resp.status_code == 200 and resp.json().get("status", False):
+            self.log(f'{resp.json().get("msg", "")} page -> {page}')
+            self.works.extend(resp.json().get("list", []))
+            if page < resp.json().get("pageInfo", {}).get("pageCount", 0):
                 self.get_study_record(course_id, page + 1)
 
     def log(self, *args, **kwargs) -> None:
@@ -414,6 +524,7 @@ class CKWK:
             self.online()
 
     def run(self) -> None:
+        print("当前为刷课模式: ")
         while self.login():
             self.get_user()
             thread = Thread(target=self.start_loop, daemon=True)
@@ -443,7 +554,7 @@ class CKWK:
                             f'{l.get("name", "")} --> {l.get("id", "")} --> 已经学习过了'
                         )
                         continue
-                    
+
                     if isReply:
                         self.reply(c.get("id", ""), l.get("id", ""))
 
@@ -470,6 +581,48 @@ class CKWK:
             self.log("所有课程已经结束")
             break
 
+    def work(self):
+        print("当前为刷作业模式: ")
+        while self.login():
+            self.get_user()
+            thread = Thread(target=self.start_loop, daemon=True)
+            thread.start()
+            cs = self.user.get("courses", [])
+            print(
+                f'姓名: {self.user.get("name", "")}\n\t互评: {self.user.get("huping", 0)}\n\t乐学园: {self.user.get("lexueyuan", 0)}\n\t讨论主题: {self.user.get("taolunzhuti", 0)}\n\t学习时长: {self.user.get("study_time", "")}'
+            )
+            if len(cs) > 1:
+                for i, _c in enumerate(cs):
+                    print(f'{i+1}: {_c.get("name", "")} {_c.get("progress", "")}')
+                cs_order = input("请输入刷课顺序(用英文逗号分隔,不需要耍的不用加): ")
+                cs_list = [cs[int(_c) - 1] for _c in cs_order.split(",")]
+            cs_list = cs
+            for c in cs_list:
+                self.get_work_record(c.get("id", ""), "")
+                # print(self.works)
+                for work in self.works:
+                    self.qas = []
+                    if "已阅" in work.get("state"):
+                        self.log(
+                            f'{work.get("title", "")} 已阅 {work.get("finalScore", "")}'
+                        )
+                        continue
+                    if not self.get_question_answer(
+                        c.get("id", ""), work.get("nodeId", ""), work.get("id", "")
+                    ):
+                        continue
+                    for q in self.qas:
+                        # print(q)
+                        chooise = self.ai.chat(q)
+                        self.work_submit(
+                            q.get("id", ""),
+                            work.get("id", ""),
+                            chooise,
+                            int(q.get("index", 1)) < len(self.qas),
+                        )
+                    # break
+            break
+
 
 if __name__ == "__main__":
     host_server = [
@@ -486,12 +639,17 @@ if __name__ == "__main__":
         u = input("请输入账号: ")
     if p == "":
         p = input("请输入密码: ")
-    r = input("是否需要评论(不需要[0],需要[1]): ")
-    if int(r) == 1:
-        isReply = True
-    else:
-        isReply = False
-    print(bool(isReply))
-    if u is None and p is None and c in [0, 1, 2]:
-        os._exit(0)
-    CKWK(f"{u}", f"{p}", host_server[c][0]).run()
+
+    # 刷课
+    # r = input("是否需要评论(不需要[0],需要[1]): ")
+    # if int(r) == 1:
+    #     isReply = True
+    # else:
+    #     isReply = False
+    # print(bool(isReply))
+    # if u is None and p is None and c in [0, 1, 2]:
+    #     os._exit(0)
+    # CKWK(f"{u}", f"{p}", host_server[c][0]).run()
+
+    # 刷作业
+    CKWK(f"{u}", f"{p}", host_server[c][0]).work()
diff --git a/pyproject.toml b/pyproject.toml
index cafa223..c1053d9 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -8,5 +8,6 @@ dependencies = [
     "ddddocr>=1.5.6",
     "dotenv>=0.9.9",
     "lxml>=6.0.2",
+    "openai>=2.8.1",
     "requests>=2.32.5",
 ]