Integrations
LangGraph migration guide
LangGraph's interrupt() gives you the primitive - a way to pause a graph node and wait for human input. Cheqpoint gives you everything built on top: notification routing, approval UI, timeout handling, audit trail, Slack/email integration, and a full dashboard. Replace weeks of infrastructure with a single function call.
Before - raw LangGraph
Weeks of infrastructure to build yourself
# 1. Configure checkpointing
config = {"interrupts": ["approval_node"]}
# 2. Define interrupt node
def approval_node(state):
raise Interrupt("Amount > £100, approve?")
# 3. Build notification system ← weeks
# 4. Build approval UI ← weeks
# 5. Handle timeouts ← days
# 6. Write to audit log ← days
# 7. Resume on approval
graph.invoke(
None,
Command(resume=True),
config
)
# Still need:
# - Slack integration
# - Dashboard
# - Escalation rules
# - Retry logic
# - State persistenceAfter - Cheqpoint
One line. Ships today.
import cheqpoint
# Exact same semantics as interrupt()
# - but with everything else included
result = cheqpoint.request(
action="issue_refund",
amount=200,
currency="GBP",
customer="james@example.com",
context="SLA breach - 4hr outage",
notify=[
"slack:#approvals",
"email:ops@company.com"
]
)
if result.approved:
execute_refund()
# Notification ✓ UI ✓
# Timeout ✓ Audit ✓
# Slack ✓ Dashboard ✓What you get out of the box
Notification routing
Slack, email, Teams, PagerDuty - all configurable per request or globally.
Approval UI
A full dashboard inbox. Reviewers see the action, context, and AI reasoning in one view.
Timeout handling
Set expiry per request. On timeout: auto-approve, auto-reject, or escalate.
Full audit trail
Every decision, who made it, when - automatically logged and exportable.
Slack interactive buttons
[Approve] / [Decline] buttons directly in Slack. No dashboard visit needed.
Analytics dashboard
Track approval rates, response times, rule matches, and agent volume.
Ready to migrate?
It takes about 10 minutes to replace your LangGraph interrupt() with Cheqpoint.