I built a mock API tool because I was tired of frontend work blocking on a backend that didn't exist yet. The design choice that mattered most: the data is seeded, not random.
Pick a dataset (say, users) and a seed, and you get exactly the same rows every time: same names, same emails, same IDs. That matters because:
- Tests that assert on specific values don't break on every run
- Snapshot tests of rendered output stay stable
- Changing the seed gives you a different but equally consistent dataset
On top of that is a real in-memory router, not canned responses. GET /api/mock/users?country=Germany&sort=-createdAt really filters and sorts the rows, with page/limit/cursor pagination, search, field selection and nested relations like /users/:id/posts.
Other pieces: mock JWT/API-key/basic auth, error simulation (400-503 or random) and latency presets for testing loading and failure states, GraphQL schema browsing, OpenAPI generation, code snippets in 15 languages, and mock WebSocket/SSE streams.
The part that trips people up: there are two modes. Simulate runs entirely in your browser (offline, zero network calls). Live endpoint hits a hosted URL with CORS enabled, which you need if a separate app or a teammate must call it.
Try: curl "https://forgeplug.com/api/mock/users?limit=5"
Tool: https://www.forgeplug.com/tools/mock-api-studio
Do you seed your mock data, or generate it fresh per run? I'd like to hear what breaks in your tests.

