Composio Cold Email: Use FoxReach as a Composio Tool

What Composio is

Composio is an agent tool platform - a catalog of pre-integrated SaaS tools that your AI agent can call through one unified interface. Instead of writing custom wrappers for Gmail + Slack + FoxReach + Notion + Linear, you authenticate each once in the Composio dashboard, and every agent framework reads them from the same catalog.

For cold email specifically, Composio makes sense when FoxReach is one tool among many in a larger agent workflow. If your agent needs to read a Linear ticket, query a Postgres database, send a Slack update, AND run a cold email campaign, Composio collapses that from 4 separate integrations into 4 entries in one catalog.

Prerequisites

• A Composio account

  Free tier available at composio.dev. Generate an API key from the dashboard.

• A FoxReach API key

  Free plan works. Settings -> API Keys in FoxReach.

• Python 3.10+ or TypeScript

  Composio ships SDKs for both. Examples below use Python.

• An agent framework

  LangChain, CrewAI, OpenAI Agents SDK, or any Composio-supported framework.

Connect FoxReach to Composio

Two steps. Authenticate once in the Composio dashboard; reference the integration from your agent code.

1. In the Composio dashboard, find FoxReach in the tool catalog. Click Connect.
2. Paste your FoxReach API key. Composio stores it encrypted and uses it whenever an agent calls a FoxReach tool.
3. Note the connected account ID - your agents reference this to identify which FoxReach workspace to use.

pip install composio-core

from composio import ComposioToolSet, App

toolset = ComposioToolSet()
# List FoxReach actions Composio knows about
tools = toolset.get_tools(apps=[App.FOXREACH])
for t in tools:
 print(t.name, "-", t.description[:80])

Use from LangChain

Composio wraps each FoxReach action as a LangChain tool. Pass them straight into any agent that accepts LangChain tools.

from composio_langchain import ComposioToolSet, App
from langchain.agents import AgentExecutor, create_tool_calling_agent
from langchain_openai import ChatOpenAI
from langchain_core.prompts import ChatPromptTemplate

toolset = ComposioToolSet()
tools = toolset.get_tools(apps=[App.FOXREACH])

llm = ChatOpenAI(model="gpt-4o-mini", temperature=0)
prompt = ChatPromptTemplate.from_messages([
 ("system", "You manage cold email campaigns. Use FoxReach tools from Composio."),
 ("human", "{input}"),
 ("placeholder", "{agent_scratchpad}"),
])
agent = create_tool_calling_agent(llm, tools, prompt)
executor = AgentExecutor(agent=agent, tools=tools, verbose=True)

executor.invoke({
 "input": "Create a draft campaign called 'Design agencies Q2' and add a 3-step sequence.",
})

Use from CrewAI

CrewAI integration follows the same pattern. Composio's crewai adapter returns CrewAI-compatible tool objects you attach to an agent's tools list.

from composio_crewai import ComposioToolSet, App
from crewai import Agent, Task, Crew

toolset = ComposioToolSet()
tools = toolset.get_tools(apps=[App.FOXREACH])

sender = Agent(
 role="Outbound Sender",
 goal="Create a draft campaign and import leads via FoxReach Composio tools.",
 backstory="You ship drafts into FoxReach. You never start a campaign - humans do that.",
 tools=tools,
 llm="gpt-4o-mini",
)

task = Task(
 description="Create a campaign called 'March founders', 3-step sequence. Draft only.",
 expected_output="campaign_id",
 agent=sender,
)

Crew(agents=[sender], tasks=[task]).kickoff()

When to use Composio vs direct MCP

Both paths work. Pick by how many tool providers your agent consumes.

For most teams building their first agent, direct FoxReach MCP is the simpler start. Swap to Composio when the agent grows into multi-tool territory and you want centralized authentication.

Common pitfalls

Frequently asked questions