LangChain

Hub: documentation home · Python API · Integrations.

In-process. No vectorsmith serve subprocess.

pip install "vectorsmith[qdrant,langchain]"

from vectorsmith import load_tools
from langchain.agents import create_agent

tools = load_tools("tools.invoices.yaml", "tools.tickets.yaml")
try:
    agent = create_agent("openai:gpt-4.1", tools)
    result = await agent.ainvoke({"messages": [{"role": "user", "content": prompt}]})
finally:
    await tools.aclose()

load_tools compiles the YAML and returns LangChain StructuredTools (requires the langchain extra / langchain-core). Mix them with your own @tools and with MCP clients (langchain-mcp-adapters) for Slack/GitHub.

Worked sample: examples/langchain_agent/.

from vectorsmith.langchain import load_tools is the same function.

LangGraph uses these tools unchanged — LangGraph.