verifiedworkflows-langchain 0.1.0

Human-in-the-loop review as agent tools for LangChain, CrewAI, LlamaIndex, and AutoGen.

verifiedworkflows-langchain

Human-in-the-loop review as agent tools for LangChain, CrewAI, LlamaIndex, and AutoGen.

Install

pip install verifiedworkflows-langchain

Quick start

import os
os.environ["VW_API_KEY"] = "vw_live_xxx"

from verifiedworkflows_langchain import get_tools
from langchain.agents import create_agent
from langchain_openai import ChatOpenAI

model = ChatOpenAI(model="gpt-4o")
tools = get_tools()

agent = create_agent(model, tools=tools)

# The agent can now submit AI output for human review
# before sending it to users
response = agent.invoke({
    "messages": [{"role": "user", "content": "Write a patient summary, then review it"}]
})

Tools

Tool	Description
submit_for_review	Submit AI output for certified human review
get_review_status	Check task status (pending, in_review, completed)
get_review_result	Get verified result with corrections and HMAC signature

Direct client (without LangChain)

from verifiedworkflows_langchain import VerifiedWorkflows

vw = VerifiedWorkflows(api_key="vw_live_xxx")
task = vw.submit(content="AI output here", review_type="medical")
result = vw.result(task["id"])  # blocks until completed

Works with

• LangChain / LangGraph — @tool decorator
• CrewAI — pass tools to agent
• LlamaIndex — FunctionTool.from_defaults
• AutoGen — pass as function tools

License

MIT