rw-check-compatibility

Check Compatibility

Analyze the user's project to determine whether it is compatible with Runway's public API.

Why This Matters

Runway's public API requires server-side invocation. The API key must never be exposed in client-side code. Projects that are purely frontend (static HTML/JS, client-only SPAs without a backend) cannot safely call the API.

Analysis Steps

Step 1: Identify the Project Type

Search the project root for these files to determine the stack:

File	Indicates
package.json	Node.js project
requirements.txt, pyproject.toml, Pipfile, setup.py	Python project
go.mod	Go project
Cargo.toml	Rust project
pom.xml, build.gradle	Java/Kotlin project
Gemfile	Ruby project
composer.json	PHP project

If none of these exist, flag the project as unknown and ask the user what language/runtime they're using.

Step 2: Check for Server-Side Capability

Look for indicators of a server/backend:

Node.js projects — check package.json dependencies for:

• express, fastify, koa, hapi, nest, hono → HTTP server framework
• next → Next.js (has API routes — compatible)
• nuxt → Nuxt.js (has server routes — compatible)
• remix → Remix (has loaders/actions — compatible)
• @sveltejs/kit → SvelteKit (has server routes — compatible)
• astro → Astro (has API endpoints if SSR enabled)

Python projects — check for:

• flask, django, fastapi, starlette, tornado, aiohttp, sanic → web server framework
• streamlit, gradio → can make server-side calls

Red flags (frontend-only):

• package.json with only react, vue, svelte, angular and NO server framework
• vite.config.ts or webpack.config.js with no server/SSR configuration
• Static site generators without server routes (e.g., plain Gatsby, plain Eleventy)
• index.html as the only entry point with inline <script> tags

Step 3: Check for Existing Runway SDK

Search for existing Runway SDK installations:

Node.js:

• Check package.json for @runwayml/sdk
• Search for import RunwayML or require('@runwayml/sdk') in source files

Python:

• Check requirements.txt / pyproject.toml for runwayml
• Search for from runwayml import RunwayML or import runwayml in source files

Step 4: Check Runtime Version

Node.js: Must be version 18 or higher (node --version) Python: Must be version 3.8 or higher (python3 --version)

Step 5: Check for Environment Variable Support

Look for .env file, .env.example, .env.local, or dotenv configuration:

• Node.js: dotenv in dependencies, or framework-native env support (Next.js, etc.)
• Python: python-dotenv in dependencies, or framework-native support

Report Format

After analysis, provide a clear report:

## Runway API Compatibility Report

**Project type:** [Node.js / Python / etc.]
**Server-side capable:** [Yes / No / Partial]
**Runtime version:** [version] — [Compatible / Needs upgrade]
**Runway SDK installed:** [Yes / No]
**Environment variable support:** [Yes / No / Needs setup]

### Verdict: [COMPATIBLE / NEEDS CHANGES / INCOMPATIBLE]

[If COMPATIBLE]
Your project is ready for Runway API integration. Proceed with API key setup.

[If NEEDS CHANGES]
Your project needs the following changes:
1. [List specific changes needed]

[If INCOMPATIBLE]
Your project is frontend-only and cannot safely call Runway's API. Options:
1. **Add a backend** — Add an Express/FastAPI server or use a framework with server routes (Next.js, SvelteKit, etc.)
2. **Use a serverless function** — Add API routes via Vercel Functions, AWS Lambda, Cloudflare Workers, etc.
3. **Create a separate backend** — Build a thin API proxy that your frontend calls

Important Notes

• Never suggest embedding the API key in client-side code. This is a security risk.
• If the project uses Next.js, Remix, SvelteKit, Nuxt, or Astro with SSR, it IS compatible — the server-side route handlers can call the API.
• Serverless platforms (Vercel, Netlify, AWS Lambda, Cloudflare Workers) are compatible.
• Docker/containerized apps are compatible if they run a server process.

After Compatibility Check

If the project is compatible, suggest the user proceed with +rw-setup-api-key to configure their API credentials.