Simulate real backends across every protocol.

Run your first mock locally without a MockForge account. Keep OpenAPI specifications and YAML configuration in Git, then add realistic data, latency, and failure behavior as your tests grow.

Get started free Star on GitHub

Run the example with Docker:

$ docker run --rm -p 3000:3000 ghcr.io/saasy-solutions/mockforge:latest serve --spec examples/openapi-demo.json --http-port 3000

In another terminal: curl http://localhost:3000/users

Measured 9.2 seconds from an uncached image pull to the first response. How we measured.

Realistic data, latency & record-replay

Local first. Configuration as code.

The open-source CLI runs your mocks on your machine or in your own CI. Your specifications and configuration are ordinary files you can review in pull requests. Hosted collaboration is optional; external AI providers and upstream integrations use the services you configure.

Start locally Configuration reference Compare development workflows

One framework, every transport your services use

REST
gRPC
GraphQL
WebSocket
Kafka
MQTT

Everything you need to fake a backend convincingly

Mock the protocol, synthesize the data, inject the failures, and inspect every request from one place.

Multi-protocol from one config

Mock HTTP REST, gRPC, GraphQL, WebSocket, Kafka and MQTT side by side. The HTTP bridge even exposes your gRPC services over REST.

REST gRPC GraphQL WebSocket Kafka MQTT

Realistic data synthesis

Template tokens, faker fields, deterministic seeding and relationship awareness.

{ "id": "{{uuid}}",
  "email": "{{faker.email}}" }

Latency & fault injection

Operation-aware latency and failure profiles so you test retries and fallbacks before production does.

OpenAPI conformance

Validate requests against your spec with deep schema resolution across path, query, header, cookie and body.

Record & replay

Capture live traffic and replay it deterministically to debug and benchmark.

A modern admin UI

Manage servers, overrides and metrics from the browser. Plugins, AI event streams and AES-256 config encryption included.

Three steps to a running mock

1
Define your API and rules

Point at an OpenAPI spec, then set latency and failure profiles per operation.
2
Generate realistic data

Use built-in tokens like {{uuid}} and {{now}}, or plug in your own generators.
3
Run and iterate

Start the server, watch requests in the admin UI, and tune behavior as you build.

$ mockforge init
✔ Created mockforge.yaml

$ mockforge add /users/{id} --status 200 \
  --body '{ "id": "{{uuid}}", "name": "{{name}}" }'

$ mockforge serve --port 8080
Listening on http://localhost:8080

Why teams adopt MockForge

Frontend and backend unblock each other with realistic mocks instead of waiting on live services, and scenario switching surfaces retry and fallback bugs before integration day.

Read the engineering notes

10+

protocols mocked from one tool

100%

open source, MIT or Apache-2.0

command to a hosted mock

real backends required in CI

Evaluating against another tool?

Side-by-side guides for where multi-protocol, realism-first workflows differ.

MockForge vs WireMock

HTTP-centric stubbing versus broader protocol coverage and realism-first simulation.

MockForge vs MockServer

Expectation-based API testing versus service virtualization and realism-driven scenarios.

MockForge vs Postman Mock

Collection-based mocks inside Postman versus multi-protocol, spec-driven simulation anywhere.

Questions, answered

Is MockForge open source?: Yes, permissively licensed under MIT or Apache-2.0. Contributions welcome.
Which protocols are supported?: REST, gRPC, GraphQL, WebSocket, Kafka and MQTT out of the box, with more via plugins.
Can I run it in CI?: Yes. Spin up MockForge alongside your tests for stable, realistic endpoints with no live backend.
Does it support record and replay?: Yes. Capture live traffic, store fixtures, and replay them deterministically.

Ready to mock any API?