How to convert Runnables as Tools

Prerequisites

This guide assumes familiarity with the following concepts:

• Runnables
• Tools
• Agents

Here we will demonstrate how to convert a LangChain Runnable into a tool that can be used by agents, chains, or chat models.

Dependencies

Note: this guide requires langchain-core >= 0.2.13. We will also use OpenAI for embeddings, but any LangChain embeddings should suffice. We will use a simple LangGraph agent for demonstration purposes.

%%capture --no-stderr
%pip install -U langchain-core langchain-openai langgraph

LangChain tools are interfaces that an agent, chain, or chat model can use to interact with the world. See here for how-to guides covering tool-calling, built-in tools, custom tools, and more information.

LangChain tools-- instances of BaseTool-- are Runnables with additional constraints that enable them to be invoked effectively by language models:

• Their inputs are constrained to be serializable, specifically strings and Python dict objects;
• They contain names and descriptions indicating how and when they should be used;
• They may contain a detailed args_schema for their arguments. That is, while a tool (as a Runnable) might accept a single dict input, the specific keys and type information needed to populate a dict should be specified in the args_schema.

Runnables that accept string or dict input can be converted to tools using the as_tool method, which allows for the specification of names, descriptions, and additional schema information for arguments.

Basic usage

With typed dict input:

from typing import List

from langchain_core.runnables import RunnableLambda
from typing_extensions import TypedDict


class Args(TypedDict):
    a: int
    b: List[int]


def f(x: Args) -> str:
    return str(x["a"] * max(x["b"]))


runnable = RunnableLambda(f)
as_tool = runnable.as_tool(
    name="My tool",
    description="Explanation of when to use tool.",
)

API Reference:RunnableLambda

print(as_tool.description)

as_tool.args_schema.schema()

Explanation of when to use tool.

{'title': 'My tool',
 'type': 'object',
 'properties': {'a': {'title': 'A', 'type': 'integer'},
  'b': {'title': 'B', 'type': 'array', 'items': {'type': 'integer'}}},
 'required': ['a', 'b']}

as_tool.invoke({"a": 3, "b": [1, 2]})

'6'

Without typing information, arg types can be specified via arg_types:

from typing import Any, Dict


def g(x: Dict[str, Any]) -> str:
    return str(x["a"] * max(x["b"]))


runnable = RunnableLambda(g)
as_tool = runnable.as_tool(
    name="My tool",
    description="Explanation of when to use tool.",
    arg_types={"a": int, "b": List[int]},
)

Alternatively, the schema can be fully specified by directly passing the desired args_schema for the tool:

from pydantic import BaseModel, Field


class GSchema(BaseModel):
    """Apply a function to an integer and list of integers."""

    a: int = Field(..., description="Integer")
    b: List[int] = Field(..., description="List of ints")


runnable = RunnableLambda(g)
as_tool = runnable.as_tool(GSchema)

String input is also supported:

def f(x: str) -> str:
    return x + "a"


def g(x: str) -> str:
    return x + "z"


runnable = RunnableLambda(f) | g
as_tool = runnable.as_tool()

as_tool.invoke("b")

'baz'

In agents

Below we will incorporate LangChain Runnables as tools in an agent application. We will demonstrate with:

• a document retriever;
• a simple RAG chain, allowing an agent to delegate relevant queries to it.

We first instantiate a chat model that supports tool calling:

OpenAI

pip install -qU langchain-openai

import getpass
import os

os.environ["OPENAI_API_KEY"] = getpass.getpass()

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(model="gpt-4o-mini")

Anthropic

pip install -qU langchain-anthropic

import getpass
import os

os.environ["ANTHROPIC_API_KEY"] = getpass.getpass()

from langchain_anthropic import ChatAnthropic

llm = ChatAnthropic(model="claude-3-5-sonnet-20240620")

Azure

pip install -qU langchain-openai

import getpass
import os

os.environ["AZURE_OPENAI_API_KEY"] = getpass.getpass()

from langchain_openai import AzureChatOpenAI

llm = AzureChatOpenAI(
    azure_endpoint=os.environ["AZURE_OPENAI_ENDPOINT"],
    azure_deployment=os.environ["AZURE_OPENAI_DEPLOYMENT_NAME"],
    openai_api_version=os.environ["AZURE_OPENAI_API_VERSION"],
)

Google

pip install -qU langchain-google-vertexai

import getpass
import os

os.environ["GOOGLE_API_KEY"] = getpass.getpass()

from langchain_google_vertexai import ChatVertexAI

llm = ChatVertexAI(model="gemini-1.5-flash")

Cohere

pip install -qU langchain-cohere

import getpass
import os

os.environ["COHERE_API_KEY"] = getpass.getpass()

from langchain_cohere import ChatCohere

llm = ChatCohere(model="command-r-plus")

NVIDIA

pip install -qU langchain-nvidia-ai-endpoints

import getpass
import os

os.environ["NVIDIA_API_KEY"] = getpass.getpass()

from langchain import ChatNVIDIA

llm = ChatNVIDIA(model="meta/llama3-70b-instruct")

FireworksAI

pip install -qU langchain-fireworks

import getpass
import os

os.environ["FIREWORKS_API_KEY"] = getpass.getpass()

from langchain_fireworks import ChatFireworks

llm = ChatFireworks(model="accounts/fireworks/models/llama-v3p1-70b-instruct")

Groq

pip install -qU langchain-groq

import getpass
import os

os.environ["GROQ_API_KEY"] = getpass.getpass()

from langchain_groq import ChatGroq

llm = ChatGroq(model="llama3-8b-8192")

MistralAI

pip install -qU langchain-mistralai

import getpass
import os

os.environ["MISTRAL_API_KEY"] = getpass.getpass()

from langchain_mistralai import ChatMistralAI

llm = ChatMistralAI(model="mistral-large-latest")

TogetherAI

pip install -qU langchain-openai

import getpass
import os

os.environ["TOGETHER_API_KEY"] = getpass.getpass()

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    base_url="https://api.together.xyz/v1",
    api_key=os.environ["TOGETHER_API_KEY"],
    model="mistralai/Mixtral-8x7B-Instruct-v0.1",
)

Following the RAG tutorial, let's first construct a retriever:

from langchain_core.documents import Document
from langchain_core.vectorstores import InMemoryVectorStore
from langchain_openai import OpenAIEmbeddings

documents = [
    Document(
        page_content="Dogs are great companions, known for their loyalty and friendliness.",
    ),
    Document(
        page_content="Cats are independent pets that often enjoy their own space.",
    ),
]

vectorstore = InMemoryVectorStore.from_documents(
    documents, embedding=OpenAIEmbeddings()
)

retriever = vectorstore.as_retriever(
    search_type="similarity",
    search_kwargs={"k": 1},
)

API Reference:Document | InMemoryVectorStore | OpenAIEmbeddings

We next create use a simple pre-built LangGraph agent and provide it the tool:

from langgraph.prebuilt import create_react_agent

tools = [
    retriever.as_tool(
        name="pet_info_retriever",
        description="Get information about pets.",
    )
]
agent = create_react_agent(llm, tools)

for chunk in agent.stream({"messages": [("human", "What are dogs known for?")]}):
    print(chunk)
    print("----")

{'agent': {'messages': [AIMessage(content='', additional_kwargs={'tool_calls': [{'id': 'call_W8cnfOjwqEn4cFcg19LN9mYD', 'function': {'arguments': '{"__arg1":"dogs"}', 'name': 'pet_info_retriever'}, 'type': 'function'}]}, response_metadata={'token_usage': {'completion_tokens': 19, 'prompt_tokens': 60, 'total_tokens': 79}, 'model_name': 'gpt-4o-mini', 'system_fingerprint': None, 'finish_reason': 'tool_calls', 'logprobs': None}, id='run-d7f81de9-1fb7-4caf-81ed-16dcdb0b2ab4-0', tool_calls=[{'name': 'pet_info_retriever', 'args': {'__arg1': 'dogs'}, 'id': 'call_W8cnfOjwqEn4cFcg19LN9mYD'}], usage_metadata={'input_tokens': 60, 'output_tokens': 19, 'total_tokens': 79})]}}
----
{'tools': {'messages': [ToolMessage(content="[Document(id='86f835fe-4bbe-4ec6-aeb4-489a8b541707', page_content='Dogs are great companions, known for their loyalty and friendliness.')]", name='pet_info_retriever', tool_call_id='call_W8cnfOjwqEn4cFcg19LN9mYD')]}}
----
{'agent': {'messages': [AIMessage(content='Dogs are known for being great companions, known for their loyalty and friendliness.', response_metadata={'token_usage': {'completion_tokens': 18, 'prompt_tokens': 134, 'total_tokens': 152}, 'model_name': 'gpt-4o-mini', 'system_fingerprint': None, 'finish_reason': 'stop', 'logprobs': None}, id='run-9ca5847a-a5eb-44c0-a774-84cc2c5bbc5b-0', usage_metadata={'input_tokens': 134, 'output_tokens': 18, 'total_tokens': 152})]}}
----

See LangSmith trace for the above run.

Going further, we can create a simple RAG chain that takes an additional parameter-- here, the "style" of the answer.

from operator import itemgetter

from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.runnables import RunnablePassthrough

system_prompt = """
You are an assistant for question-answering tasks.
Use the below context to answer the question. If
you don't know the answer, say you don't know.
Use three sentences maximum and keep the answer
concise.

Answer in the style of {answer_style}.

Question: {question}

Context: {context}
"""

prompt = ChatPromptTemplate.from_messages([("system", system_prompt)])

rag_chain = (
    {
        "context": itemgetter("question") | retriever,
        "question": itemgetter("question"),
        "answer_style": itemgetter("answer_style"),
    }
    | prompt
    | llm
    | StrOutputParser()
)

API Reference:StrOutputParser | ChatPromptTemplate | RunnablePassthrough

Note that the input schema for our chain contains the required arguments, so it converts to a tool without further specification:

rag_chain.input_schema.schema()

{'title': 'RunnableParallel<context,question,answer_style>Input',
 'type': 'object',
 'properties': {'question': {'title': 'Question'},
  'answer_style': {'title': 'Answer Style'}}}

rag_tool = rag_chain.as_tool(
    name="pet_expert",
    description="Get information about pets.",
)

Below we again invoke the agent. Note that the agent populates the required parameters in its tool_calls:

agent = create_react_agent(llm, [rag_tool])

for chunk in agent.stream(
    {"messages": [("human", "What would a pirate say dogs are known for?")]}
):
    print(chunk)
    print("----")

{'agent': {'messages': [AIMessage(content='', additional_kwargs={'tool_calls': [{'id': 'call_17iLPWvOD23zqwd1QVQ00Y63', 'function': {'arguments': '{"question":"What are dogs known for according to pirates?","answer_style":"quote"}', 'name': 'pet_expert'}, 'type': 'function'}]}, response_metadata={'token_usage': {'completion_tokens': 28, 'prompt_tokens': 59, 'total_tokens': 87}, 'model_name': 'gpt-4o-mini', 'system_fingerprint': None, 'finish_reason': 'tool_calls', 'logprobs': None}, id='run-7fef44f3-7bba-4e63-8c51-2ad9c5e65e2e-0', tool_calls=[{'name': 'pet_expert', 'args': {'question': 'What are dogs known for according to pirates?', 'answer_style': 'quote'}, 'id': 'call_17iLPWvOD23zqwd1QVQ00Y63'}], usage_metadata={'input_tokens': 59, 'output_tokens': 28, 'total_tokens': 87})]}}
----
{'tools': {'messages': [ToolMessage(content='"Dogs are known for their loyalty and friendliness, making them great companions for pirates on long sea voyages."', name='pet_expert', tool_call_id='call_17iLPWvOD23zqwd1QVQ00Y63')]}}
----
{'agent': {'messages': [AIMessage(content='According to pirates, dogs are known for their loyalty and friendliness, making them great companions for pirates on long sea voyages.', response_metadata={'token_usage': {'completion_tokens': 27, 'prompt_tokens': 119, 'total_tokens': 146}, 'model_name': 'gpt-4o-mini', 'system_fingerprint': None, 'finish_reason': 'stop', 'logprobs': None}, id='run-5a30edc3-7be0-4743-b980-ca2f8cad9b8d-0', usage_metadata={'input_tokens': 119, 'output_tokens': 27, 'total_tokens': 146})]}}
----

See LangSmith trace for the above run.