mockd is a high-performance, multi-protocol mock server that provides simulation capabilities for seven protocols in a single binary, officially positioned as "one binary, seven protocols, zero dependencies." It is designed to solve the problem that other mock tools usually support only a single protocol and require additional runtime or extensions to be installed.
Main capabilities
- Multi-protocol support: built-in support for HTTP, gRPC, GraphQL, WebSocket, MQTT, SSE, and SOAP. The official documentation provides a comparison table of protocol support and import/export capabilities, marking statuses such as built-in, requires a separate extension, paid/hosted tier only, partially implemented, or roadmap.
- Spec import and digital twins: can import real API specs (such as OpenAPI), bind mocks to stateful data tables, and build mock services that can be tested with real SDKs. Official verification results include: Stripe's 49/49 stripe-go SDK tests passing, Twilio's 13/13 twilio-go SDK tests passing, and OpenAI Python SDK (models, assistants, chat completions) verification passing. For complete digital twin configurations, see mockd-samples.
- Stateful CRUD and deterministic responses: supports stateful resource management and deterministic seeded responses, making it easy to build repeatable, verifiable test environments.
- Chaos engineering: built-in chaos injection capability to simulate failure scenarios in mocks.
- Request logs and validation: provides request log viewing, contract validation, and other features, and supports workspace management and import/export.
- Secure connections and multiplexing: all seven protocols can be multiplexed over a single secure connection on port 443, suitable for environments behind NAT and firewalls.
- Web UI: release versions provide on the management port (http://localhost:4290）提供 Web UI, including a VS Code-like editor, command palette, mock tree with folders, request log viewer, and near-miss debugging features.
AI-native (MCP) mockd has a built-in Model Context Protocol (MCP) server with 18 tools. AI agents can create mocks, manage state, import specs, and validate contracts without ever touching the CLI. It is compatible with Claude Desktop, Cursor, Windsurf, and any editor that supports MCP. The tools cover mock CRUD, stateful resources, chaos injection, request logs, validation, workspaces, and import/export.
How to use
- Download prebuilt binaries for Linux, macOS, and Windows from the GitHub Releases page.
- The official website, documentation (docs.mockd.io), examples (mockd-samples), and contribution guide (CONTRIBUTING.md) are provided.
- Fully available offline, no account required.
- For teams that need a shared environment, the official sources mention that persistent cloud mocks, team management, cloud tunnels, and other features are under development, and a waitlist is open, but specific information is subject to official future releases.
Use cases
- Backend development and integration testing that need to simulate multiple protocol APIs at the same time.
- Wanting to build digital twins based on real API specs so that mocks can pass real SDK tests.
- Needing to inject chaos, validate contracts, and record request logs in CI/CD or local development.
- Teams or individual developers who want AI agents to create and manage mocks directly in the editor via MCP.
Limitations and notes
- The free open-source version uses the Apache License 2.0 and can be used for commercial purposes free of charge.
- Features marked as "Cloud" in the official table exist only in the paid/hosted tier, and the free tier has limitations.
- WireMock import of OpenAPI and Postman import do not include community converters in the OSS standalone JAR.
- Features such as cloud mocks, team management, and cloud tunnels are in a "coming soon" state and have not yet been officially released.
- Specific installation commands, configuration details, and complete pricing information are not provided in the given materials and require reference to the official documentation.
