How agent-ready is PlanetScale?

PlanetScale achieved an overall agentability score of 34/100, indicating significant barriers for autonomous AI agents attempting to interact with its public web presence. The homepage scored 34/100, the pricing page 42/100, and documentation 28/100. Machine readability is strong at 100/100, meaning the underlying HTML structure is sound, but chunking (59/100) and control (48/100) show room for improvement.

The most critical gaps appear in transparency, status reporting, and clean handoffs—each scoring 0/100. This means agents receive no visibility into system reasoning, no feedback during asynchronous operations, and insufficient error context when requests fail. Shadow UI avoidance scored 15/100, suggesting interactive elements may not be reliably detectable or actionable by automated systems.

These scores indicate that while PlanetScale's content is technically parseable, agents will struggle to complete multi-step workflows, recover from errors, or understand the system's state and reasoning. Addressing the zero-scored principles would yield the highest impact on agent-readiness.

PlanetScale can improve agentability by implementing the following changes:

• Add structured error responses that include error codes, the specific field or resource that failed, and actionable context in both UI messages and API payloads instead of generic failure text.
• Implement visible status updates for asynchronous operations with descriptive text (e.g., 'Uploading 3 of 10 files') and wrap them in ARIA live regions (role=status or aria-live) so agents can detect progress programmatically.
• Expose an activity log or audit trail in machine-readable format (JSON event feed) that records system actions, enabling agents to verify what occurred and troubleshoot issues.
• Surface confidence indicators on key outputs—such as a confidence score in API responses or 'verified'/'best-guess' labels in the UI—so agents can assess reliability.
• Provide expandable result summaries with source references: show a one-line summary that agents can parse quickly, with links or IDs pointing to the underlying data or documentation.
• For long-running tasks, display multi-step progress indicators or activity feeds that enumerate each phase, allowing agents to track workflow state and detect stalls.