Documentation

The fastest way to use GetMocked is through the MCP server. Connect it to your AI coding assistant and create, update, and test mock APIs using natural language, without leaving your editor.

Quick Setup

npx @getmocked/mcp login

// .cursor/mcp.json
{
  "mcpServers": {
    "getmocked": {
      "command": "npx",
      "args": ["@getmocked/mcp"]
    }
  }
}

// .vscode/settings.json
{
  "mcp": {
    "servers": {
      "getmocked": {
        "command": "npx",
        "args": ["@getmocked/mcp"]
      }
    }
  }
}

Available Tools

Once connected, these tools are available to your AI assistant:

Example Workflow

With the MCP server connected, you can talk to your AI assistant naturally:

Every mock API gets a unique endpoint URL and API key. Use them like any real API.

Making Requests

curl -X POST https://api.getmocked.ai/m/{userId}/{slug} \
  -H "Authorization: Bearer {your-api-key}" \
  -H "Content-Type: application/json" \
  -d '{"name": "Example input"}'

Both POST and GET are supported. For GET requests, pass input as query parameters.

Response

Responses are AI-generated JSON that matches your output schema. The data is realistic and contextually appropriate — not just random values.

Project APIs

APIs inside a project use a slightly different URL:

https://api.getmocked.ai/m/{userId}/{projectSlug}/{apiSlug}

Find your exact endpoint URL and API key on each API's Deploy tab in the dashboard.

Test how your app handles errors by adding query parameters to any mock API request.

Combine them for complex scenarios: ?_mock_status=503&_mock_delay=1500 simulates a slow 503 error.

Delay without a status code returns a normal successful response after the wait — useful for testing loading states and spinners.

Error responses include "simulated": true so you can distinguish them programmatically.

Quality Modes

Control the quality and speed of generated mock data.

Set via the X-Mock-Mode header or ?_mode=fast query param. You can also set a default mode per API in the dashboard. Free tier is limited to fast mode.

Caching

GetMocked caches generated responses for 7 days. Use the cache to get instant responses and consistent data across requests.

Set via the X-Mock-Cache header or ?_cache=any query param. Cached responses don't count as generation requests (no AI cost).

Need responses that stick around forever? See Data Persistence for permanent, input-matched storage.

Data persistence gives your mock APIs sticky responses — enable it on any API and the same input will always return the same output. Persisted responses are stored permanently, cost $0 to serve, and never expire.

How It Works

Controlling Persistence

Override the default behavior per-request with the X-Mock-Persist header or ?_persist query param:

Response Source

Every response includes an X-Mock-Source header so you know where the data came from:

Persistence vs Caching

Managing Persisted Data

View, edit, and delete persisted entries from the Data tab on any API in the dashboard. You can also manage data programmatically:

# List persisted entries
curl https://api.getmocked.ai/api/mock-apis/{id}/persisted-data \
  -H "Authorization: Bearer {token}"

# Delete a specific entry
curl -X DELETE https://api.getmocked.ai/api/mock-apis/{id}/persisted-data/{inputHash} \
  -H "Authorization: Bearer {token}"

# Clear all persisted data
curl -X DELETE https://api.getmocked.ai/api/mock-apis/{id}/persisted-data \
  -H "Authorization: Bearer {token}"

Rate Limits

When you hit your limit, the API returns 429 Too Many Requests with a reset_at timestamp telling you when your quota refreshes.