Skip to content

Introduction

Testzilla runs automated tests against your voice AI agents (Retell, VAPI, and more), evaluates the calls, and tracks quality over time. You can drive everything from the web app, the REST API, or an MCP server for AI agents like Claude Code and Cursor.

Projects

Top-level containers that group related channels and tests.

Channels

A connection to an agent under test on a provider (Retell / VAPI), including the phone number and platform config.

Tests

A scenario the AI tester runs against a channel, with a prompt, persona, and pass/fail criteria.

Test Results

The outcome of a run: transcript, recording, evaluation, latency, and pass/fail.

REST API

Automate projects, channels, tests, and results.

Head to the Quickstart to make your first API call.