Skip to main content

Quick Start

Add a trust layer to your existing Temporal agent with OpenBox. This guide assumes you already have a working Temporal agent and shows you how to wrap it with OpenBox for trust, monitoring, and compliance.

Building from Scratch?

If you don't have a Temporal agent yet, see Temporal Integration Guide for a complete end-to-end setup.

Prerequisites​

  • Existing Temporal agent with workflows and activities
  • Python 3.10+ installed
  • OpenBox Account - Sign up at app.openbox.ai

Step 1: Register Your Agent in OpenBox​

Before wrapping your worker, create an agent in the OpenBox platform:

  1. Log in to the OpenBox Dashboard
  2. Navigate to Agents → Click Add Agent
  3. Configure the agent:
    • Workflow Engine: Temporal
    • Agent Name: Your agent name (e.g., "Customer Support Agent")
    • Agent ID: Auto-generated
    • Description: What your agent does
    • Teams: assign the agent to one or more teams
    • Icon: select an icon
  4. API Key Generation:
    • Click Generate API Key
    • Copy and store the key (shown only once)
  5. Configure platform settings:
    • Initial Risk Assessment (AIVSS) - select a risk profile (Level 1-4)
    • Attestation (Execution Evidence) - select AWS KMS
    • Goal Alignment (drift detection) - set an alignment threshold and drift action
  6. Click Add Agent

See Registering Agents for a field-by-field walkthrough of the form.

tip

Your API key format: obx_live_xxxxxxxxxxxxxxxxxx

Step 2: Install OpenBox SDK​

Add the OpenBox SDK to your existing project:

pip install openbox-temporal-sdk-python

# Or with uv
uv add openbox-temporal-sdk-python

Step 3: Configure Environment Variables​

Add OpenBox credentials to your environment:

# .env or export these
OPENBOX_URL=https://api.openbox.ai
OPENBOX_API_KEY=obx_live_your_api_key_here

Step 4: Wrap Your Existing Worker​

Replace your standard Temporal Worker with OpenBox's create_openbox_worker:

Before (Standard Temporal):

from temporalio.client import Client
from temporalio.worker import Worker

async def main():
    client = await Client.connect("localhost:7233")
    
    worker = Worker(
        client,
        task_queue="my-task-queue",
        workflows=[MyWorkflow],
        activities=[my_activity],
    )
    
    await worker.run()

After (With OpenBox):

import os
from temporalio.client import Client
from openbox import create_openbox_worker  # Changed import

async def main():
    client = await Client.connect("localhost:7233")
    
    # Replace Worker with create_openbox_worker
    worker = create_openbox_worker(
        client=client,
        task_queue="my-task-queue",
        workflows=[MyWorkflow],
        activities=[my_activity],
        
        # Add OpenBox configuration
        openbox_url=os.getenv("OPENBOX_URL"),
        openbox_api_key=os.getenv("OPENBOX_API_KEY"),
    )
    
    await worker.run()

That's it! Your workflows and activities remain completely unchanged.

Step 5: Run Your Worker​

Start your worker as usual:

python worker.py

Your agent now runs with the OpenBox trust layer enabled.

Step 6: Verify in OpenBox Dashboard​

Trigger a workflow and view it in the dashboard:

  1. Run a workflow (using your existing trigger/client code)
  2. Open the OpenBox Dashboard
  3. Navigate to Agents → Click your agent
  4. Go to Monitor tab to see:
    • Active and completed sessions
    • Session replay with full event timeline
    • Captured HTTP requests (LLM calls, API requests)
    • Activity inputs/outputs
    • Governance decisions

What OpenBox Captures​

The SDK automatically sends these events to OpenBox:

  • Workflow events: Started, completed, failed, signals
  • Activity events: Started (with input), completed (with output), duration
  • HTTP telemetry: Request/response bodies, headers, status codes
  • Database operations (optional): SQL queries, NoSQL operations

All captured data is evaluated against your governance policies in real-time.

Next Steps​

Now that your agent is running with OpenBox:

  1. Configure Trust Controls - Set up guardrails, policies, and behavioral rules
  2. Monitor Sessions - Use Session Replay to debug and audit agent behavior
  3. Set Up Approvals - Add human-in-the-loop for sensitive operations
  4. Advanced Configuration - Fine-tune timeouts, fail policies, and event filtering

Need More Details?​