keploy/keploy

What it solves

Keploy eliminates the manual effort of writing API and integration tests by automatically generating them from real user traffic. It solves the problem of achieving high test coverage (up to 90%) without requiring developers to write extensive boilerplate code or manually create complex data mocks for external dependencies.

How it works

Keploy uses eBPF to capture network traffic at the network layer, making it code-less and language-agnostic. It records API calls, database queries, and streaming events (like Kafka or RabbitMQ) and then replays them as deterministic tests. It also uses AI to analyze existing recordings and OpenAPI schemas to identify boundary values and missing fields to expand API coverage.

Who it’s for

Developers and QA engineers who need to create comprehensive integration tests for any language or framework, and those who whom want to avoid the manual provisioning of infrastructure for testing.

Highlights

• No Code Changes: Captures traffic via eBPF, requiring no SDKs or code modifications.
• Infra-Virtualization: Mocks not just HTTP endpoints, but also databases (Postgres, MySQL, MongoDB) and message queues (Kafka, RabbitMQ).
• AI-Driven Expansion: Uses AI to find boundary values and missing fields to increase test coverage.
• Language Agnostic: Works with any runtime or framework because it operates at the network layer.
• CI/CD Integration: Supports running tests locally, in Jenkins, GitHub Actions, or Kubernetes clusters.