Open-source API and integration testing platform that auto-generates tests and mocks from real user traffic using eBPF.
Keploy is an open-source platform for API and integration testing that automatically generates tests and data mocks by recording real user traffic. It captures API calls, database queries, and streaming events using eBPF at the network layer, then replays them as deterministic tests without requiring code changes. It solves the problem of creating comprehensive, isolated test environments and achieving high test coverage quickly.
Developers and QA engineers building and testing distributed applications, microservices, or APIs who need automated, language-agnostic testing with high coverage. It's especially useful for teams practicing CI/CD and DevOps.
Developers choose Keploy because it generates tests and mocks directly from production traffic with no code changes, supports complex infra virtualization (databases, queues), provides combined coverage metrics, and integrates seamlessly into CI/CD pipelines. Its eBPF-based approach makes it uniquely language-agnostic and non-invasive.
Open-source platform for creating safe, isolated production sandboxes for API, integration, and E2E testing.
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Uses eBPF at the network layer, requiring no SDKs or code changes, making it compatible with any programming language or framework as highlighted in the README.
Records and replays databases, message queues, and external APIs beyond HTTP, enabling deterministic tests without real infrastructure dependencies.
Converts real API calls into tests and mocks automatically, reducing manual effort and achieving high coverage faster than traditional unit tests.
Provides combined reports for statement, branch, API schema, and business use-case coverage, offering insights for both developers and QA teams.
The README admits that some dependencies with non-open-source protocols are not supported, which can hinder adoption in enterprise environments with proprietary systems.
Since tests are generated from recorded traffic, edge cases not encountered during recording may be missed, requiring supplemental manual testing or AI expansion, which isn't foolproof.
Requires eBPF capabilities and proper kernel permissions, which can complicate setup in restricted, legacy, or certain cloud environments, adding operational overhead.