feat: add MiniMax M2.7 as LLM provider

README.md

@@ -74,6 +74,7 @@ Create **.env** file under the repo root directory with your own API keys:
 ```
 OPENAI_API_KEY=sk-proj-...
 DEEPSEEK_API_KEY=sk-...
+MINIMAX_API_KEY=your-minimax-api-key
 ```
 
 ## Record
@@ -141,6 +142,16 @@ python run_replay.py --model-provider openai --wap_replay_list data_processed/ex
 ```
 For **smart-replay**, replace the path with a smart‑replay JSON to test this mode.
 
+Supported `--model-provider` values: `azure`, `anthropic`, `openai`, `ollama`, `minimax`.
+
+To use **MiniMax** (M2.7, 204K context):
+```bash
+# Set MINIMAX_API_KEY in your .env file, then run:
+python run_replay.py --model-provider minimax --wap_replay_list data_processed/smart_replay/wap_smart_replay_list_<task_id>.json --max-concurrent 1
+```
+
+To use MiniMax in the **MCP client** (`mcp_client.py`), set `LLM_PROVIDER=minimax` and `MINIMAX_API_KEY` in your `.env` file.
+
 ## Convert to MCP Server
 
 ```bash

mcp_client.py

@@ -225,6 +225,12 @@ class LLMClient:
 
     def __init__(self, api_key: str) -> None:
         self.api_key: str = api_key
+        self.provider: str = os.getenv("LLM_PROVIDER", "openai").lower()
+        # Resolve MiniMax key at init so tests can patch the environment once
+        if self.provider == "minimax":
+            self._minimax_key: str = os.getenv("MINIMAX_API_KEY", api_key)
+        else:
+            self._minimax_key = ""
 
     def get_response(self, messages: list[dict[str, str]]) -> str:
         """Get a response from the LLM.
@@ -238,15 +244,22 @@ def get_response(self, messages: list[dict[str, str]]) -> str:
         Raises:
             httpx.RequestError: If the request to the LLM fails.
         """
-        url = "https://api.openai.com/v1/chat/completions"
+        if self.provider == "minimax":
+            url = "https://api.minimax.io/v1/chat/completions"
+            api_key = self._minimax_key
+            model = "MiniMax-M2.7"
+        else:
+            url = "https://api.openai.com/v1/chat/completions"
+            api_key = self.api_key
+            model = "gpt-4o"
 
         headers = {
             "Content-Type": "application/json",
-            "Authorization": f"Bearer {self.api_key}",
+            "Authorization": f"Bearer {api_key}",
         }
         payload = {
             "messages": messages,
-            "model": "gpt-4o"
+            "model": model
         }
 
         try:

run_replay.py

@@ -112,6 +112,14 @@ def get_llm_model_generator(
         elif model_provider == "ollama":
             llm = ChatOllama(model="ota-preview-v16", num_ctx=20000,temperature=0)
             yield llm
+        elif model_provider == "minimax":
+            llm = ChatOpenAI(
+                model="MiniMax-M2.7",
+                temperature=0.01,
+                openai_api_key=os.getenv("MINIMAX_API_KEY", ""),
+                openai_api_base="https://api.minimax.io/v1",
+            )
+            yield llm
         else:
             raise ValueError(f"Invalid model provider: {model_provider}")
 
@@ -258,6 +266,7 @@ async def process_with_semaphore(
                 "anthropic",
                 "openai",
                 "ollama",
+                "minimax",
             ],
         )
         parser.add_argument(