Cloudinary for Builders & Agents

Pick your stack. Cloudinary is already in the context.

MCP-ready inside Lovable's chat, with snippets sized for agent-first app flows: upload, transform, deliver.

Native MCP connector and embedded media workflows for Bolt-generated apps.

Hosted MCP and copy-paste snippets that fit Base44's no-NPM install flow.

Cursor rules + Cloudinary MCP help your AI pair-programmer wire media handling into your app.

Cloudinary skill + MCP server. Upload, transform, and deliver media in the same session you build.

Hook Cloudinary into the Codex agent — uploads, transforms, and search available in the same session it writes code.

Cloudinary nodes for n8n workflows — automate ingestion, transformation, tagging, and delivery without writing pipeline code.

Official Shopify integration for syncing product imagery and serving optimized storefront media.

Go beyond next/image with uploads, transformations, and optimized delivery for real app media.

Build media-rich interfaces with Cloudinary URLs, SDKs, and optimized assets ready for every screen.

Handle uploads, transformations, and delivery from your backend without building a custom media pipeline.

Deploy on Vercel. Use Cloudinary for the media layer: upload, transform, optimize, deliver.

Run on AWS, deliver through Cloudinary. Drop in alongside S3 or Lambda — Cloudinary handles transform and delivery on top.

Run on Google Cloud, deliver through Cloudinary. Same SDKs, same MCP — works alongside GCS and Cloud Run.

From the first upload, every asset is something your agent can operate on.

Media is not a flat file system. It's a library your product is built on. The builders who understand this build better products, faster — storefronts that load, UGC that doesn't break, and video that just works — without a media rewrite at month six.

Addressable, not just stored.Every asset gets a stable public ID, structured metadata, and a permanent URL. Your code refers to an asset by what it is — not by where it lives in a folder.
Searchable by what's in it.Auto-tagged on upload — objects, scenes, dominant colors, OCR'd text. Query by content through a search API, not by guessing the filename someone uploaded six months ago.
Structured on the way in.Uploads are validated, deduplicated, and routed by preset — formats normalized, duplicates collapsed, folders enforced. The library stays clean even when 50 sources are writing into it.
Versioned, not overwritten.Originals are immutable. Every transform — crop, resize, format — is a derivative computed on demand. Roll back, regenerate, or restore without losing what you started with.

terminal

// Your agent uploads the product photo.
const asset = await cld.uploader.upload("hero.jpg", {
  folder: "products",
  tags: ["hero", "ai-generated"],
  context: { campaign: "launch-q3" },
  categorization: "google_tagging",
});

// Later — any agent, any time — finds it back.
const matches = await cld.search
  .expression("tags:hero AND context.campaign=launch-q3")
  .execute();

Your media is a product feature.
Time to treat it like one.

Build an ecommerce storefront

Product shots, variants, and thumbnails that load fast on every device and surface — from the first catalog upload.

catalogvariantsdelivery

Add video without a video team

Upload once, deliver everywhere — desktop, mobile, low-bandwidth — without building a transcoding pipeline.

videoencodingstreaming

Accept user and vendor uploads

Messy inputs come in. Clean assets come out — moderated, organized, transformed, ready to ship.

uploadsmoderationnormalize

Ship to every channel

Turn one source asset into the right format, size, and crop for product pages, social, marketplaces, email, and more…

socialmarketplaceemail

What your agent gets with Cloudinary.

Intent-aware delivery

Tell Cloudinary the destination — product page, marketplace tile, Instagram story, email header — and get back the right asset. We hold the spec. As platforms change their rules, your code doesn't.

productmarketplacesocialemailweb

AI-native transformations

Background removal, smart crop, generative fill, generative replace, AI upscaling, auto-captioning — one API call each. No model orchestration to write.

bg_removalgen_fillsmart_cropupscale

Global delivery & governance

Auto-optimized for every device, format, and network — AVIF, WebP, responsive variants on by default. Brand controls, audit logs, and asset-level permissions ready when your team is.

f_autoq_autoresponsiveaudit

The “we'll add real media later” trap.

Most AI agents will tell your founder to start with the default image component and add a real media layer "later." That advice is cheap on day one and expensive at month six.

Day one

Images in a flat folder. Your agent can upload, but can't find.

Month 3

Optimization, variants, delivery — wired by hand, one at a time.

Month 6

Migration. While the app is already live.

From sign-up to ship-ready
in five minutes.

01 — Sign up

Get an API key.

Free tier, no card. Or create a dedicated agent key — its own permissions, rate limits, and audit log.

Get API Key - Free ↗

02 — Connect

Install the MCP server.

One command. Claude Code, Cursor, Bolt, your own — your agent runtime discovers all Cloudinary tools automatically.

03 — Build

Your agent handles media.

Upload, organize, transform, search, deliver. You ship the app — no pipeline code, no CDN config, no spec-chasing.

terminal

// transform + deliver
const url = cld.image("hero")
  .resize(fill().width(1200))
  .delivery(format("auto"))
  .delivery(quality("auto"))
  .toURL();

View SDK docs →

The same platform 11,000+ brands run on — now wired for agents.

Trusted at scale by 11,000+ brands. Now callable by agents.

Ship your first optimized asset
in under five minutes.

Get a free API key, install the SDK, and deliver your first transformed image before your prototype becomes a media mess.

01 - Get an API key

Free tier. No credit card required.

02 - Install the SDK

One package. Works with the stack you already use.

03 - Transform and deliver

Resize, optimize, and serve through Cloudinary.

Get API Key - Free ↗

What builders are using it for

AI app builder

Vibe-coded apps with images

Building a media-heavy app in Bolt, Lovable, or Cursor? Add Cloudinary from day one. Free tier, instant setup, no migration later. Your LLM knows how to use it - Cloudinary is already in the context.

Agentic pipeline

Automated content operations

Agents that ingest, transform, tag, and deliver media assets at scale - across regions, formats, and channels - without a human approving every step. Cloudinary is the media layer those pipelines run on.

Claude / GPT skill

Media handling as an agent skill

Wire Cloudinary into your Claude Code or GPT agent as a native skill. Your agent uploads, transforms, and retrieves assets in the same session it builds the rest of the app. No context switching.

Enterprise

AI-driven creative operations

Campaign asset generation, personalization at scale, global delivery - orchestrated by agents, powered by Cloudinary. Scoped credentials, audit logs, and usage analytics built in for enterprise governance.

Your agent builds the app.
Cloudinary handles the media.

Your agent builds the app.
Cloudinary handles the media.

Pick your stack. Cloudinary is already in the context.

From the first upload, every asset is something your agent can operate on.

Your media is a product feature.
Time to treat it like one.