langgraph-approval-hub 0.2.0

Python SDK for LangGraph Approval Hub — human-in-the-loop approval workflows

langgraph-approval-hub

Python SDK for LangGraph Approval Hub — add human-in-the-loop approval workflows to your LangGraph agents in 3 lines.

Install

pip install langgraph-approval-hub

Usage

from langgraph_approval_hub import request_approval

decision = request_approval(
    hub_url="https://your-hub.vercel.app",
    api_token="your-api-secret-token",
    agent_name="Finance Agent",
    action_description="Process $4,200 refund batch for 12 customers",
    assignee="finance-team",
    assignee_type="team",
)

if decision == "approved":
    process_refunds()

request_approval() blocks until a human approves or rejects via the dashboard, then returns "approved" or "rejected".

Parameters

Parameter	Type	Required	Default	Description
hub_url	str	✅	—	Your Approval Hub URL
api_token	str	✅	—	Your API_SECRET_TOKEN env var value
agent_name	str	✅	—	Name shown in the dashboard
action_description	str	✅	—	Plain-English description of what the agent wants to do
assignee	str	✅	—	Email address or team name
assignee_type	str	✅	—	"person" or "team"
agent_reasoning	str	❌	None	Step-by-step reasoning shown to approver
escalate_to	str	❌	None	Email to escalate to if timeout exceeded
timeout_minutes	int	❌	60	Minutes before escalation
poll_interval	int	❌	10	Seconds between status polls

Links

• Live Demo
• GitHub
• Deploy Your Own (Free)