Stability AI

Getting Started

Stability AI provides a powerful suite of generative AI models, most notably Stable Diffusion, that allow you to create images, videos, audio, and 3D assets from text prompts. This guide is for developers, designers, and content creators who want to integrate Stability AI's capabilities into their workflows or applications using the API.

Setup & Registration

1. Create an Account – Go to platform.stability.ai and sign up with your email or a Google/Apple account. No credit card is required for the initial sign-up.
2. Verify Your Email – Check your inbox and click the verification link. This is immediate.
3. Get Your API Key – After logging in, navigate to the API Keys section in your dashboard. Click Create API Key, give it a name, and copy the key. Store it securely – you will not be able to see it again.
4. Free Credits – New accounts receive a grant of free credits (typically 25 generations worth of standard images). You can monitor your usage on the Billing page.

Your First API Call

The following curl command generates an image using the Stable Diffusion 3.5 Large model. Replace YOUR_API_KEY with the key you just created.

curl -X POST https://api.stability.ai/v2beta/stable-image/generate/sd3 \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Accept: image/*" \
  -F "prompt=A serene mountain lake at sunset, photorealistic" \
  -F "output_format=jpeg"

What it does

• Sends a POST request to the /generate/sd3 endpoint.
• The Authorization header includes your API key.
• The -F flags send multipart form data: the text prompt and desired output format.
• The response is a binary JPEG image (which you can redirect to a file).

Expected output

A high-quality JPEG image of a mountain lake at sunset. If you want to see the JSON response with base64 data, remove the Accept: image/* header and add -F "accept=application/json".

# Save the image to a file
curl -o output.jpg ...

Common Use Cases

1. Marketing Visuals

Generate product photos or social media graphics in seconds. For example, create a sleek image of a coffee mug on a wooden table with a specific lighting style.

curl -X POST ... \
  -F "prompt=Modern coffee mug on rustic wooden table, soft morning light, product photography"

2. Design Ideation

Quickly iterate on mood boards, character concepts, or architectural sketches. Use negative prompts to exclude unwanted elements.

curl -X POST ... \
  -F "prompt=futuristic city skyline, cyberpunk style" \
  -F "negative_prompt=blurry, low quality, cartoon"

3. Game Asset Creation

Generate backgrounds, textures, or character sprites for game development. Use the aspect_ratio parameter to match your canvas size.

curl -X POST ... \
  -F "prompt=ancient stone temple interior, video game background, 16:9" \
  -F "aspect_ratio=16:9"

Tips & Best Practices

Stay Within Free Tier Limits

• Each image generation costs 1 credit (or more for higher resolution / larger models). Monitor your remaining credits on the dashboard.
• Use lower resolution (e.g., 512x512 instead of 1024x1024) to stretch credits.
• Avoid generating many images in a short burst, as rate limits apply (5 requests per second on the free tier).

Common Pitfalls to Avoid

• Expired or missing API key – Always double-check that your key is correctly copied and not expired. Regenerate it if needed.
• Incorrect model name – The endpoint URL includes the model version (e.g., sd3 for Stable Diffusion 3.5). Sending to a wrong endpoint returns a 404.
• Prompt too vague – Include specific style, lighting, and composition keywords. Use negative prompts to reduce artifacts.

Performance Optimization

• For bulk generation, use the async endpoint (/v2beta/stable-image/generate/sd3/async) to avoid timeouts.
• If you are a developer, consider using the official Python SDK (pip install stability-sdk) for cleaner code and built-in retry logic.
• Cache generated images locally to avoid re-generating the same prompt.

By following these steps, you can quickly start creating professional-grade visuals with Stability AI.