Introduction

Workspace Quickstart in 3 Steps

1. Open the dashboard and enter the API key from Polar checkout.
2. Create a client workflow profile and generate QA scenarios.
3. Evaluate a transcript or webhook endpoint, then create a shareable report.

What AgentPrey Is

AgentPrey is now centered on workflow QA for agencies shipping AI receptionists and lead-intake automation. The web workspace models client business rules, generates scenario coverage, evaluates conversations, and produces report-ready evidence.

What It Does

• Stores client workflow profiles behind API-key access.
• Generates scenario suites from local-service templates.
• Evaluates pasted transcripts or bounded webhook endpoint runs.
• Stores transcripts, failures, evidence, risk scores, and suggested fixes.
• Creates public-by-link QA reports for client delivery.
• Keeps the open-source CLI scanner and uploaded scan reports available as legacy-supported surfaces.

Who It Is For

• AI automation agencies testing local-service receptionists before client launch.
• Operators who need repeatable lead-intake regression checks.
• Teams that still need the legacy prompt-injection CLI scanner for local or CI security scans.

Next Steps

Start with QA Workspace. For CLI scanning, use Legacy CLI Quickstart.

Current Limitations

• API-key login is the MVP auth model.
• Webhook runs are bounded to 10 scenarios and use redacted endpoint metadata.
• No team accounts, native voice calls, file uploads, or enterprise auth in this phase.
• The Rust CLI remains supported but is no longer the primary product surface.