Documentation / Performance and Scaling

Performance and Scaling

Updated February 20, 2026

Impact on Page Load Times

Agent Builder has zero impact on front-end page load times for your visitors. The plugin loads its assets only in the WordPress admin dashboard and on pages where you have specifically embedded a chat widget via shortcode.

The AI processing happens server-side through API calls to your AI provider. These calls only occur when you (or a scheduled task) actively interacts with an assistant.

Server Resource Usage

  • Memory: Minimal — the plugin itself uses very little PHP memory. AI API calls are handled as standard HTTP requests.
  • CPU: Negligible for normal usage. The heavy processing happens on the AI provider’s servers, not yours.
  • Database: Audit logs are stored as custom post types. On an active site, this might add a few hundred rows per month. Logs can be cleaned up periodically.
  • Disk space: The plugin is under 5 MB. File backups (from assistant-modified files) may use additional space.

Hosting Requirements

Agent Builder runs well on standard WordPress hosting. Specific requirements:

  • Shared hosting: Works fine for normal usage (a few conversations per day)
  • Managed WordPress hosting (e.g., WP Engine, Kinsta): Ideal — reliable cron and good PHP performance
  • VPS/Dedicated: Recommended for heavy usage with scheduled tasks and event automation

Scaling Considerations

  • Concurrent users: Each chat session creates API calls to your AI provider. If multiple admins chat simultaneously, costs increase linearly.
  • Scheduled tasks: Run sequentially through WordPress cron. They do not strain your server.
  • Event automation: Async processing means events queue up and process one at a time — no server overload.
  • Multisite: Each site in a multisite network has its own assistant configuration and API key.

Optimization Tips

  • Use a server-level cron job instead of WP-Cron for more reliable scheduling
  • Review and clean the audit log periodically (Agent Builder → Audit Log)
  • Deactivate assistants you are not using to reduce background processing
  • Set spending limits in your AI provider dashboard to control costs
Browse all documentation →