- Updated: March 22, 2026
- 2 min read
Building a Workflow‑Automation AI Agent with the OpenClaw Full‑Stack Template
Guide: Building a Workflow‑Automation AI Agent with OpenClaw
In this tutorial we walk developers through creating a powerful workflow‑automation AI agent using the OpenClaw full‑stack template. The steps cover environment setup, defining workflow steps, integrating external services, and one‑click deployment on UBOS.
1. Environment Setup
Start by installing the UBOS CLI and cloning the OpenClaw template:
curl -sSL https://ubos.tech/install.sh | bash
ubos init openclaw-agent
cd openclaw-agent
Install the required dependencies:
npm install
pip install -r requirements.txt
2. Defining Workflow Steps
The OpenClaw template ships with a workflow.yaml file. Define each step of your AI agent, for example:
steps:
- name: fetch-data
action: http.get
url: https://api.example.com/data
- name: process-data
action: python.run
script: scripts/process.py
- name: generate-response
action: openai.chat
model: gpt-4
3. Integrating External Services
Use the built‑in integrations/ folder to add connectors. For a Slack notification add:
integrations/slack.yaml:
webhook_url: "YOUR_SLACK_WEBHOOK"
channel: "#alerts"
Reference the integration in a workflow step:
- name: notify-slack
action: slack.post_message
config: integrations/slack.yaml
4. One‑Click Deployment on UBOS
When the workflow is ready, commit your changes and push to the UBOS registry:
git add .
git commit -m "Add workflow"
git push ubos main
UBOS will automatically build and deploy the agent. You can monitor the deployment from the UBOS dashboard.
5. Publish Your Agent
After deployment, expose the agent via a public endpoint. The UBOS platform provides a secure URL you can share with users.
For a deeper dive into hosting OpenClaw on UBOS, see the dedicated guide: Host OpenClaw on UBOS.
Congratulations! You now have a fully functional workflow‑automation AI agent running on UBOS.