DummyJSON MCP. Mock every API call and data flow without a live backend.

DummyJSON gives your AI agent a massive sandbox to play in. You're not hooking up to a real backend, so you don't sweat setting up databases or worrying about live data integrity. This server lets your agent run any kind of CRUD operation—create, read, update, delete—against simulated structured data for testing and quality assurance.

It’s instant mocking for whatever API interaction you need."

"Your AI client handles the heavy lifting by calling specific tools. Here's a breakdown of what your agent can do with this pool of fake data:

User Accounts and Authentication Flow

It starts with user profiles. Your agent runs auth_login to simulate logging in, which returns dummy JWT access and refresh tokens, letting you test session-dependent calls like auth_get_me, which retrieves the profile for the logged-in user. To manage accounts generally, your agent can create a new profile using add_user, update existing details with update_user (like changing an email or name), and delete an account entirely via delete_user.

For directory testing, it runs list_users to get the full roster of every profile, or uses get_user and filter_users to pull specific user data based on criteria. Furthermore, your agent can see all shopping carts linked to a particular person using get_user_carts."

"### Product Catalog and E-commerce

Testing an e-commerce flow is simple. Your agent runs list_products for every item in the catalog or uses search_products to find goods by name, description, or keyword. For targeted testing, it pulls all items using get_products_by_category, which requires a category ID. The lifecycle of any product can be tested: your agent creates listings with add_product, fetches full details with get_product, modifies prices or descriptions using update_product, and simulates removal via delete_product.

You'll also find tools to manage categories, running list_categories for a complete listing or just list_category_list to get the simple names."

"### Content Management: Posts and Comments

This server lets you simulate an entire blogging platform. To view content, your agent runs list_posts to get every post entry available or uses search_posts to filter posts using keywords or specific criteria. You can fetch all details for one piece with get_post, and if the user is associated with a post, they can see it via get_user_posts.

The commenting system is robust: your agent retrieves every comment stored in the system by calling list_comments, or gets the full content of one specific comment using get_comment and its ID. When testing related posts, it calls get_comments_by_post to get all comments for a given post ID, or uses get_post_comments to retrieve those connected to a specific post object.

Creating new content involves running add_post, which adds an entirely new entry, or using update_post to edit the content and metadata of one that's already out there. For cleanup, it runs delete_post to permanently remove a piece of content."

"### Recipes and Cooking Data

If you’re building anything food-related, this works. Your agent can run list_recipes to see every recipe in the database or use search_recipes if it needs to find recipes based on keywords or ingredients. For specific filtering, it retrieves all meals matching a type (like dinner) using get_recipes_by_meal_type, or filters by required tags with get_recipes_by_tag.

The recipe lifecycle includes adding new ones with add_recipe, fetching full details via get_recipe, modifying ingredients through update_recipe, and removing them with delete_recipe. It also helps manage the taxonomy, running list_recipe_tags to see what tags are available."

"### Shopping Carts and To-Do Lists

For e-commerce transactions, your agent can run list_carts to get a global view of all carts. It creates a new cart with add_cart, reads the details of one specific cart using get_cart, modifies items or quantities in an existing cart via update_cart, and wipes it clean entirely with delete_cart. For task management, your agent runs list_todos to get every item.

It adds a new task with add_todo, fetches details for one specific todo using get_todo, marks things as done or changes due dates by running update_todo, and clears items out with delete_todo."

"### Quotes, Search, and Random Data

For quick content feeds, your agent pulls a single random quote using get_random_quote, or grabs up to ten for bulk testing by running get_random_quotes. It can list every stored quote with list_quotes or get one specific quote detail using get_quote. The system also offers general search capability: run search_users to find people matching criteria, or use the dedicated search_products and search_recipes functions for deep searches.

Finally, it includes helper tools like get_random_todo, which pulls a single random task item.