An installer CLI for an AI-assisted onboarding skill that guides you through integrating Cloudinary into a new or existing project.
Install the CLI globally or in your project:
npx @cloudinary-devs/cloudinary-get-started
This installs the skill to .claude/skills/cloudinary-get-started by default.
To install to a different IDE's skill folder:
npx @cloudinary-devs/cloudinary-get-started --target cursor
npx @cloudinary-devs/cloudinary-get-started --target agents
Once installed, run the onboarding guide in any project:
/cloudinary-get-started
The skill walks through seven steps:
- Silent explore — Detect your project structure, framework, and stack (automatically)
- Stage 1: AI tooling — Install Cloudinary MCP servers (required) and optionally the related skills
- Stage 2: Framework detection — Identify your stack (Django, Rails, Next.js, etc.) and deployment model (front-end, back-end, or full-stack)
- Stage 3: SDK setup — Install the official Cloudinary SDK for your framework and create
.env.example - Stage 4: Credentials — Retrieve your Cloudinary account credentials and save them to
.env - Stage 5: Validation — Verify media delivery, measure optimization savings, and document your setup
- What's next — Get personalized next steps, install the VS Code extension, and access use-case-specific prompts
Cloudinary MCP servers:
cloudinary-asset-mgmt— Query and manage assets in your Cloudinary accountcloudinary-env-config— Read and manage environment variables
Related skills (optional, install separately):
cloudinary-docs— Answers Cloudinary questions using real, up-to-date documentationcloudinary-transformations— Generates valid image and video transformation URLs that follow best practicescloudinary-react— Provides React SDK patterns, configuration, and troubleshooting (used only if you choose React)
See the main skills repo to install the related skills.
| File | Purpose |
|---|---|
bin/index.js |
Installer CLI — copies the skill into your project's IDE folder |
skill/SKILL.md |
Skill definition — stage order, gates, and stage completion format |
skill/cloudinary-get-started-full-prompt.md |
Full onboarding prompt (copy/paste into your AI assistant) |
skill/references/ |
Implementation details for each stage |
When the skill runs, it creates or updates these files in your project root:
- MCP server config — location depends on your IDE/agent, for example:
- Claude Code:
.mcp.json - Cursor:
.cursor/mcp.json - VS Code:
.vscode/mcp.json
- Claude Code:
- SDK configuration — updated in your app's entry point (e.g.,
app.py,server.js,app/config.rb) with Cloudinary imports and config - Dependency manifest —
requirements.txt/package.json/Gemfile/go.mod/ etc. — updated with Cloudinary SDK .env.example— placeholder credentials (safe to commit).env— created if it doesn't exist, or updated only with your permission if it already exists; contains real credentials (added to.gitignore, never commit)docs/cloudinary-environment.json— validation results and setup details (no secrets)docs/cloudinary-getting-started-preview.html— side-by-side image preview (front-end and full-stack projects only)