# Force tool calling in langraph swarm **URL:** \ **Category:** LangGraph\ **Tags:** python-help\ **Created:** [October 1, 2025, 11:30pm UTC](https://forum.langchain.com/t/force-tool-calling-in-langraph-swarm/1726 "2025-10-01T23:30:25Z")\ **Posts on this page:** 10\ **Page:** 2
**Author:** ![polo15s](https://yyz1.discourse-cdn.com/flex007/user_avatar/forum.langchain.com/polo15s/32/1161_2.png) [@polo15s](https://forum.langchain.com/u/polo15s)\ **Post date:** [October 2, 2025, 5:39pm UTC](https://forum.langchain.com/t/force-tool-calling-in-langraph-swarm/1726/21 "2025-10-02T17:39:36Z")
Sorry that wont be possible. I have signed a NDA and its a entire product and the company wont allow to share it ---
**Author:** ![pawel-twardziak](https://yyz1.discourse-cdn.com/flex007/user_avatar/forum.langchain.com/pawel-twardziak/32/960_2.png) [@pawel-twardziak](https://forum.langchain.com/u/pawel-twardziak)\ **Post date:** [October 2, 2025, 5:41pm UTC](https://forum.langchain.com/t/force-tool-calling-in-langraph-swarm/1726/22 "2025-10-02T17:41:41Z")
Ok, I understand. Then tell me the exact versions (langchain, langgraph) you are using - I’ll try to reproduce the code to some extent. ---
**Author:** ![pawel-twardziak](https://yyz1.discourse-cdn.com/flex007/user_avatar/forum.langchain.com/pawel-twardziak/32/960_2.png) [@pawel-twardziak](https://forum.langchain.com/u/pawel-twardziak)\ **Post date:** [October 2, 2025, 6:58pm UTC](https://forum.langchain.com/t/force-tool-calling-in-langraph-swarm/1726/23 "2025-10-02T18:58:52Z")
btw this code works for me: ```auto import os from typing import Optional from dotenv import load_dotenv from langchain_core.messages import HumanMessage from langgraph.prebuilt import create_react_agent from langchain.chat_models import init_chat_model from langgraph_swarm import create_swarm from langchain_core.tools import tool load_dotenv(verbose=True) # -----------------------llm---------------------- llm = init_chat_model( "claude-3-7-sonnet-latest", model_provider="anthropic", api_key=os.getenv("ANTHROPIC_API_KEY", ""), ) @tool() def document_retriever(query: str, domain_name: str, user_folder: Optional[str] = None, website_name: Optional[str] = None) -> str: """ Tool to retrieve relevant documents using hybrid search (vector + BM25) from web-scraped content. Args: query (str): The search query domain_name (str): Name of the domain/website for the vector store user_folder (str, Optional): Username to use in the path (optional) website_name (str, Optional): Name of the website folder (optional, defaults to "brandsmithworldwide") Returns: Retrieved content as formatted text """ print(f"Retrieving `{query}` from `{domain_name}` (user: `{user_folder})`, website: `{website_name})`") return "His name is Paul." def make_rag_prompt(): return "You are an helpful assistant." def get_all_agent_tools(): return [document_retriever] def create_rag_agent(): """Create the RAG agent with all tools from ToolFactory""" # Get all tools for the agent all_tools = get_all_agent_tools() # llm_with_tools = llm.bind_tools(all_tools, strict=True) print('--' * 30) print(f"Creating RAG agent with {len(all_tools)} tools") for _tool in all_tools: tool_name = getattr(_tool, 'name', 'unnamed_tool') print(f"Tool: {tool_name}") print('--' * 30) # model_with_tools = self.model.bind_tools(tools=all_tools, tool_choice={"type": "tool", "name": "document_retriever"}) return create_react_agent( model=llm.bind_tools(all_tools, tool_choice="auto"), tools=all_tools, prompt=make_rag_prompt(), name="rag_agent" ) builder = create_swarm( [create_rag_agent()], default_active_agent="rag_agent" ) graph = builder.compile() answer = graph.invoke({"messages": [HumanMessage("Hi! What is the name of the main actor of \"No way!\" movie? Domain: movies. Folder: all. Website: NONE.")]}) print(answer["messages"][1]) ``` Response: ```terminal ------------------------------------------------------------ Creating RAG agent with 1 tools Tool: document_retriever ------------------------------------------------------------ Retrieving `No way! movie main actor` from `movies` (user: `all)`, website: `None)` content=[{'text': 'To find information about the main actor of the "No way!" movie, I can search for this in the movies domain. Let me do that for you.', 'type': 'text'}, {'id': 'toolu_01X4MuruiDVbcth48Jii2Nim', 'input': {'query': 'No way! movie main actor', 'domain_name': 'movies', 'user_folder': 'all'}, 'name': 'document_retriever', 'type': 'tool_use'}] additional_kwargs={} response_metadata={'id': 'msg_019SnJfmkSLWy9q3EPHbAU78', 'model': 'claude-3-7-sonnet-20250219', 'stop_reason': 'tool_use', 'stop_sequence': None, 'usage': {'cache_creation': {'ephemeral_1h_input_tokens': 0, 'ephemeral_5m_input_tokens': 0}, 'cache_creation_input_tokens': 0, 'cache_read_input_tokens': 0, 'input_tokens': 601, 'output_tokens': 130, 'server_tool_use': None, 'service_tier': 'standard'}, 'model_name': 'claude-3-7-sonnet-20250219'} name='rag_agent' id='run--5a5c2b5d-20a2-4816-8f3d-b691c2432957-0' tool_calls=[{'name': 'document_retriever', 'args': {'query': 'No way! movie main actor', 'domain_name': 'movies', 'user_folder': 'all'}, 'id': 'toolu_01X4MuruiDVbcth48Jii2Nim', 'type': 'tool_call'}] usage_metadata={'input_tokens': 601, 'output_tokens': 130, 'total_tokens': 731, 'input_token_details': {'cache_read': 0, 'cache_creation': 0, 'ephemeral_5m_input_tokens': 0, 'ephemeral_1h_input_tokens': 0}} ``` ---
**Author:** ![pawel-twardziak](https://yyz1.discourse-cdn.com/flex007/user_avatar/forum.langchain.com/pawel-twardziak/32/960_2.png) [@pawel-twardziak](https://forum.langchain.com/u/pawel-twardziak)\ **Post date:** [October 2, 2025, 7:01pm UTC](https://forum.langchain.com/t/force-tool-calling-in-langraph-swarm/1726/24 "2025-10-02T19:01:19Z")
When I remove `Return` from docstring it still works: ``` """ Tool to retrieve relevant documents using hybrid search (vector + BM25) from web-scraped content. Args: query (str): The search query domain_name (str): Name of the domain/website for the vector store user_folder (str, Optional): Username to use in the path (optional) website_name (str, Optional): Name of the website folder (optional, defaults to "brandsmithworldwide") """ ``` Response: ```auto ------------------------------------------------------------ Creating RAG agent with 1 tools Tool: document_retriever ------------------------------------------------------------ Retrieving `No way! movie main actor` from `movies` (user: `all)`, website: `None)` content=[{'text': 'I\'ll help you find information about the main actor of the "No way!" movie. Let me search for that information using the document retriever tool.', 'type': 'text'}, {'id': 'toolu_01Bd5jGKwwFbSGsvFtzC1D5V', 'input': {'query': 'No way! movie main actor', 'domain_name': 'movies', 'user_folder': 'all'}, 'name': 'document_retriever', 'type': 'tool_use'}] additional_kwargs={} response_metadata={'id': 'msg_01GJaPJaUNYufnd99h6WrHwR', 'model': 'claude-3-7-sonnet-20250219', 'stop_reason': 'tool_use', 'stop_sequence': None, 'usage': {'cache_creation': {'ephemeral_1h_input_tokens': 0, 'ephemeral_5m_input_tokens': 0}, 'cache_creation_input_tokens': 0, 'cache_read_input_tokens': 0, 'input_tokens': 586, 'output_tokens': 129, 'server_tool_use': None, 'service_tier': 'standard'}, 'model_name': 'claude-3-7-sonnet-20250219'} name='rag_agent' id='run--c77b9e43-2bf1-40fe-92a9-9459e838a36c-0' tool_calls=[{'name': 'document_retriever', 'args': {'query': 'No way! movie main actor', 'domain_name': 'movies', 'user_folder': 'all'}, 'id': 'toolu_01Bd5jGKwwFbSGsvFtzC1D5V', 'type': 'tool_call'}] usage_metadata={'input_tokens': 586, 'output_tokens': 129, 'total_tokens': 715, 'input_token_details': {'cache_read': 0, 'cache_creation': 0, 'ephemeral_5m_input_tokens': 0, 'ephemeral_1h_input_tokens': 0}} ``` ---
**Author:** ![polo15s](https://yyz1.discourse-cdn.com/flex007/user_avatar/forum.langchain.com/polo15s/32/1161_2.png) [@polo15s](https://forum.langchain.com/u/polo15s)\ **Post date:** [October 4, 2025, 11:56pm UTC](https://forum.langchain.com/t/force-tool-calling-in-langraph-swarm/1726/25 "2025-10-04T23:56:16Z")
Extactly! I know there is no issue with the code, yet i get the error. I dont know how to fix it ---
**Author:** ![pawel-twardziak](https://yyz1.discourse-cdn.com/flex007/user_avatar/forum.langchain.com/pawel-twardziak/32/960_2.png) [@pawel-twardziak](https://forum.langchain.com/u/pawel-twardziak)\ **Post date:** [October 5, 2025, 10:12am UTC](https://forum.langchain.com/t/force-tool-calling-in-langraph-swarm/1726/26 "2025-10-05T10:12:46Z")
Could it be magic? 🙂 There must be a reason somewhere 🧐 ---
**Author:** ![heisenberg-7](https://avatars.discourse-cdn.com/v4/letter/h/73ab20/32.png) [@heisenberg-7](https://forum.langchain.com/u/heisenberg-7)\ **Post date:** [October 5, 2025, 1:12pm UTC](https://forum.langchain.com/t/force-tool-calling-in-langraph-swarm/1726/27 "2025-10-05T13:12:01Z")
> [@pawel-twardziak](#): > > exact versions (langchain, langgraph) you are using @polo15s Could you share the versions to try to reproduce on those exact version? ---
**Author:** ![polo15s](https://yyz1.discourse-cdn.com/flex007/user_avatar/forum.langchain.com/polo15s/32/1161_2.png) [@polo15s](https://forum.langchain.com/u/polo15s)\ **Post date:** [October 6, 2025, 7:21am UTC](https://forum.langchain.com/t/force-tool-calling-in-langraph-swarm/1726/28 "2025-10-06T07:21:36Z")
> [@pawel-twardziak](#): > > `bind_tools(all_tools, tool_choice="auto"),` Hey, I solved the issue The issue was with the wrapper function in the class. ---
**Author:** ![heisenberg-7](https://avatars.discourse-cdn.com/v4/letter/h/73ab20/32.png) [@heisenberg-7](https://forum.langchain.com/u/heisenberg-7)\ **Post date:** [October 6, 2025, 8:55am UTC](https://forum.langchain.com/t/force-tool-calling-in-langraph-swarm/1726/29 "2025-10-06T08:55:01Z")
What was the issue? ---
**Author:** ![pawel-twardziak](https://yyz1.discourse-cdn.com/flex007/user_avatar/forum.langchain.com/pawel-twardziak/32/960_2.png) [@pawel-twardziak](https://forum.langchain.com/u/pawel-twardziak)\ **Post date:** [October 6, 2025, 11:13am UTC](https://forum.langchain.com/t/force-tool-calling-in-langraph-swarm/1726/30 "2025-10-06T11:13:52Z")
@polo15s it would be great if you share the solution code so that it benefits others 🙂 [Previous page](https://forum.langchain.com/t/force-tool-calling-in-langraph-swarm/1726.md?page=1)