Skip to content

The agent engineering platform.

feat(anthropic): auto-apply mcp beta header (#34301)

Mason Daugherty3b7abdff9609
diff --git a/libs/partners/anthropic/langchain_anthropic/chat_models.py b/libs/partners/anthropic/langchain_anthropic/chat_models.py
index 29494212bc..78d5cfa94a 100644
--- a/libs/partners/anthropic/langchain_anthropic/chat_models.py
+++ b/libs/partners/anthropic/langchain_anthropic/chat_models.py
@@ -141,6 +141,7 @@ _TOOL_TYPE_TO_BETA: dict[str, str] = {
"web_fetch_20250910": "web-fetch-2025-09-10",
"code_execution_20250522": "code-execution-2025-05-22",
"code_execution_20250825": "code-execution-2025-08-25",
+ "mcp_toolset": "mcp-client-2025-11-20",
"memory_20250818": "context-management-2025-06-27",
"computer_20250124": "computer-use-2025-01-24",
"computer_20251124": "computer-use-2025-11-24",
@@ -1611,14 +1612,20 @@ class ChatAnthropic(BaseChatModel):
"type": "url",
"url": "https://docs.langchain.com/mcp",
"name": "LangChain Docs",
- # "tool_configuration": { # optional configuration
- # "enabled": True,
- # "allowed_tools": ["ask_question"],
- # },
# "authorization_token": "PLACEHOLDER", # optional authorization
}
]
+ # Optional: configure which tools are enabled via mcp_toolset
+ tools = [
+ {
+ "type": "mcp_toolset",
+ "mcp_server_name": "LangChain Docs",
+ # "default_config": {"enabled": False}, # disable all by default
+ # "configs": {"ask_question": {"enabled": True}}, # allowlist
+ }
+ ]
+
model = ChatAnthropic(
model="claude-sonnet-4-5-20250929",
mcp_servers=mcp_servers,
@@ -1626,7 +1633,7 @@ class ChatAnthropic(BaseChatModel):
response = model.invoke(
"What are LangChain content blocks?",
- tools=[{"type": "mcp_toolset", "mcp_server_name": "LangChain Docs"}],
+ tools=tools,
)
```