Deploy Your Claude Code Agent

Connect your GitHub repo, deploy agents, schedule runs, and trigger via webhooks

DataGen Agents lets you deploy Claude Code agents from your GitHub repositories. Once deployed, you can schedule them to run automatically or trigger them via webhooks.

Why Deploy Your Agent?

You already built the agent. It works. But running it locally means it’s isolated from your systems—it can’t respond to events, run on a schedule, or integrate with your workflows.

Deployed agents become part of your infrastructure:

• Seamless integration - Your agent becomes a service that other systems can call
• Event-driven execution - React instantly to webhooks from Slack, GitHub, or any external system
• Scheduled automation - Daily CRM hygiene, weekly reports, hourly monitoring—all hands-free

Why Deploy with DataGen?

Deploying Claude Agent SDK yourself means configuring container sandboxes, managing MCP OAuth flows manually, and handling long-running process infrastructure. DataGen handles all of this - connect your GitHub repo and deploy with one click.

---

Prerequisites

Create an Agent in Claude Code

Use the /agents command in Claude Code to create your agent. This creates an agent file at .claude/agents/your-agent.md.

See the Claude Code sub-agents documentation for details on creating agents.

Add MCP Capabilities (Optional)

Want your agent to use MCP tools like Google, LinkedIn, or other integrations?

We currently only support MCP through DataGen. To add MCP capabilities to your agent, add your MCP servers through DataGen. This way, your MCP tools work both locally during development and in your deployed agent.

Push Your Repo to GitHub

Your agent code needs to be in a GitHub repository.

1. Create a new repository on GitHub:

• Go to github.com/new
• Enter a repository name
• Leave “Initialize this repository with a README” unchecked
• Click Create repository

2. Get your remote origin URL:

• After creating the repo, GitHub shows the remote URL
• Copy the HTTPS URL: https://github.com/your-username/your-repo.git
• Or use SSH: [e-mail]:your-username/your-repo.git

3. Push your local code:

git init
git add .
git commit -m "Initial commit"
git remote add origin https://github.com/your-username/your-repo.git
git push -u origin main

See the GitHub documentation for more details.

---

Step 1: Add DataGen GitHub App to Your Repo

1. Go to Agents Page

   Navigate to the Agents page in DataGen.

2. Click Connect Repository

   Click the Connect Repository button to start the setup process.

   
3. Install GitHub App

   Click Install GitHub App to authorize DataGen to access your repositories.

4. Select Account or Organization

   Choose the GitHub account or organization where your agent repository lives.

5. Choose Repositories

   Select the specific repositories you want to connect, or grant access to all repositories.

6. Return to DataGen

   After authorization, return to DataGen and select your repository from the list.

---

Step 2: Deploy an Agent

1. Select Your Agent

   Click on the agent you want to deploy from the list. DataGen automatically discovers agents from your .claude/agents/ directory.

   
2. Configure Secrets

   Select which secrets your agent needs access to. You need either an Anthropic API key or a Claude Code subscription token.

   Option A: Anthropic API Key

   Add your ANTHROPIC_API_KEY as a secret in DataGen Secrets.

   Option B: Claude Code Subscription (Recommended)

   If you’re a Claude Code subscriber, you can use your subscription instead of an API key:

   1. Run claude setup-token in your terminal to generate an OAuth token
   2. Create a secret named CLAUDE_CODE_OAUTH_TOKEN in DataGen Secrets
   3. Paste your token as the secret value
   4. Select this secret when deploying your agent

   
3. Toggle Create Pull Request (Optional)

   Enable Create Pull Request if you want your agent to submit changes as PRs instead of committing directly.

4. Click Deploy Agent

   Click Deploy Agent to create your deployment.

5. Copy Your Webhook URL

   After deployment, copy the webhook URL. You’ll use this to trigger your agent externally.

   

---

Step 3: Trigger Your Agent

Once deployed, you have three ways to trigger your agent. Choose the method that fits your use case.

Schedule

Set up automatic recurring runs for your deployed agent.

1. Open Schedule Settings

   Click Schedule on your deployed agent.

   
2. Choose Schedule Type

   Select your preferred schedule:

   • Hourly - Run every hour
   • Daily - Run once per day at a specific time
   • Weekly - Run on specific days of the week
   • Custom - Define a custom cron expression
3. Set Timezone

   Choose your timezone to ensure the agent runs at the expected times.

4. Add Payload (Optional)

   Provide optional JSON payload that will be passed to your agent on each run.

   {
     "target_branch": "main",
     "dry_run": false
   }
   
5. Save Schedule

   Click Save to activate the schedule.

Run Now

Trigger your agent immediately for testing or one-off executions.

1. Click Run Now

   Click the Run Now button on your deployed agent.

   
2. Add Custom Payload (Optional)

   Optionally provide a JSON payload for this specific run.

   {
     "task": "Review latest PRs",
     "notify": true
   }
   
3. Start Execution

   Click Run to start the agent execution.

4. View Logs

   Monitor the execution in real-time through the logs view.

Webhook

Integrate your agent with external systems using the webhook URL.

Copy Your Webhook URL

Find the webhook URL on your deployed agent’s detail page.

Example: Trigger with cURL

curl -X POST "https://api.datagen.dev/agents/webhook/YOUR_WEBHOOK_ID" \
  -H "Content-Type: application/json" \
  -d '{"task": "Process new data", "source": "external"}'

Integration Examples

Use your webhook URL with:

• n8n - Include in automation workflows
• HeyReach - Real-time LinkedIn campaign monitoring
• Fireflies - Auto-generate RFPs after meetings

---

Step 4: View Execution History & Logs

Finding Your Logs

Click on a deployed agent to view its execution history. Each run shows:

• Start time - When the execution began
• Duration - How long the run took
• Status - Current state of the execution
• Trigger - What initiated the run (schedule, manual, webhook)

Understanding Status

Status	Description
Queued	Run is waiting to start
Running	Agent is currently executing
Completed	Run finished successfully
Failed	Run encountered an error

Viewing Pull Requests

If your agent creates PRs, you’ll see a link to the PR directly in the execution details.

---

Troubleshooting

Agent not appearing after connecting repo

Make sure your agent file is located at .claude/agents/your-agent.md in your repository and has been pushed to the main branch.

Webhook returns 401 Unauthorized

Verify you’re using the correct webhook URL. Each deployed agent has a unique webhook endpoint.

Scheduled run didn't execute

Check that your schedule is active and the timezone is set correctly. View the execution history for any failed runs.

Agent fails with API key error

Ensure you’ve configured the required secrets (especially your Anthropic API key) and selected them when deploying the agent.