Smart: An Automated API Testing Platform for Generating Test Cases from Manual Testing

1. Problem Background

QA teams using a generic automation framework faced high data preparation costs, environment provisioning overhead, and XML‑based case authoring complexity, especially for database preparation and cleanup.

2. Feature Overview

Smart provides a new way to write automation cases by recording manual testing behavior and data. The workflow includes:

Manual testing on Smart (similar to Postman).

Automatic generation of automation cases by reusing the current environment and data.

Execution of generated cases for regression testing, with export capabilities for integration with existing frameworks.

3. Feature Demonstration

Steps to use Smart:

Add the target API to Smart's interface list.

Use the "Interface Debug" feature to manually invoke the API.

After a successful manual test, click "Start Recording" with the request parameters and database info to generate an automation case.

Smart records the API calls and database interactions, then presents the generated case details.

4. Implementation Details

The core consists of an "Interface Recording Module" and a "Database Recording Module".

Interface Recording: Captures request parameters, invokes the target service, and records responses to generate input data and assertions.

Database Recording: Listens to MySQL binlog and general log to capture data changes, producing backup, cleanup, prepare_data , and assert_data for the case.

To ensure accurate database recording, Smart integrates with QTracer to obtain the full trace ID and filter out unrelated "noise" data. For extracting tables involved in complex SQL statements, Smart leverages the open‑source tool Inception to parse the SQL syntax tree and retrieve the table set.

5. Summary

Smart was officially released in September 2017, supporting HTTP, Dubbo, and MySQL interfaces, and can generate automation cases within minutes, dramatically reducing case authoring time. It is being adopted across multiple departments, with plans to enhance scenario coverage, functionality, and usability, and to strengthen case management capabilities.