memecub

How can I get memes in the right size through an API or MCP server?

Use the memecub REST API for search (1 credit per call), or connect the remote MCP server at https://memecub.com/mcp to Claude, ChatGPT or Cursor so your agent can search and pick memes itself. The API returns the meme image, and you resize on your side to the platform target: 1080×1080 for square feeds, 1080×1920 for stories and reels, 1080×1350 for portrait posts. A keyless trial gives you 20 searches a day, so you can test sizing before committing to a plan.

How the memecub API and MCP server work

The REST API is a search endpoint: you send a query, it returns matching memes, and each call costs 1 credit. You then take the image URL and resize or crop it to your target dimensions before posting.

The remote MCP server lives at https://memecub.com/mcp. Add it as an MCP server in Claude, ChatGPT, Cursor or another MCP-capable client, and the model can search the library on its own when it needs a meme for a message or a paragraph.

There is also a CLI, npx memecub, and a skill you install with npx skills add https://memecub.com into Claude Code, Codex, Cursor and similar tools. The skill teaches the agent when and how to search.

No key is needed to try it: the keyless trial allows 20 searches a day. Paid plans start at free with 1000 credits a month, Hobby at $3/mo, Pro at $9/mo and Scale at $29/mo, with credit packs from $5.

Sizing the image after you fetch it

The API gives you the meme; the size is your job. Pick the ratio first, then crop or pad to the exact pixels.

Square feed posts: 1080×1080. This is the safest default and what most platforms still favor for feed images.

Stories and reels: 1080×1920, a 9:16 ratio. Keep captions away from the top and bottom edges where UI overlays sit.

Portrait feed: 1080×1350, a 4:5 ratio, which takes more vertical space than a square without being cropped by the feed.

Landscape: 1200×628 for link previews and 16:9 for video thumbnails. If you only remember one number, remember 1080 pixels wide.

A minimal flow: search, download the original, resize to the target with your image library, then post. Do not upscale a small source; pick a larger original instead.

Choosing between API, MCP, CLI and skill

Use the REST API when you are building a product feature and want explicit control over queries, caching and resizing in your own code.

Use the MCP server when the consumer is a chat assistant or coding agent that should decide on its own which meme fits a message. It is the least code.

Use the CLI or skill when you work inside an editor or terminal and want the agent to fetch memes without wiring a server config by hand.

If nothing in the library fits, the AI meme maker at /generate creates one from a character, action and caption, about 10 seconds per image at 500 credits each, and refuses real people. The AI meme placement tool at /match takes an article and picks a meme for each paragraph.

Tone, licensing and etiquette

Most memes in the library are collected from the web: fine for personal chats, copyright stays with the original creators, and not for commercial use.

If you need commercial use, turn on the filter that shows only openly licensed images from Openverse, open-source emoji and public-domain art.

When an agent picks memes automatically, review the output before publishing. A meme that fits the paragraph but not the audience is worse than no meme.

Let your AI find the memes

REST API, remote MCP, CLI and a skill. Try 20 searches a day without a key.

More that fit

FAQ

What is a good size for a meme?

1080×1080 pixels for square feed posts is the safe default. For stories and reels use 1080×1920, and for portrait feed posts 1080×1350.

Is 4:3 or 16:9 better for social media?

Neither is the best default. Feed images work best near 1:1 or 4:5, while 16:9 suits video thumbnails and link previews. Use 4:3 only when the source image is already that shape.

How do I set up a meme MCP server for Claude, ChatGPT or Cursor?

Add https://memecub.com/mcp as a remote MCP server in your client. The model can then search the library itself; the keyless trial allows 20 searches a day.

Is there a meme API that returns images sized for social media?

The memecub API returns the meme image and charges 1 credit per search; you resize to the platform target on your side. That keeps the source quality intact and lets you output several sizes from one fetch.

How do I get my AI agent to find the right meme for a message?

Give it the MCP server or install the skill with `npx skills add https://memecub.com`, then let it search by meaning rather than keyword. For longer text, the /match tool picks a meme per paragraph.