Skip to main content
View as MarkdownOpen in ChatGPTOpen in Claude

Stability AI Provider

The current Go provider exposes Stability text-to-image generation through the shared ai.GenerateImage API.

Setup​

Installation​

import (
    "context"
    "github.com/digitallysavvy/go-ai/pkg/ai"
    "github.com/digitallysavvy/go-ai/pkg/providers/stability"
)

Configuration​

provider := stability.New(stability.Config{
    APIKey: os.Getenv("STABILITY_API_KEY"),
})

model, err := provider.ImageModel("stable-diffusion-xl-1024-v1-0")

Get API Key​

export STABILITY_API_KEY=sk-...

Available Models​

provider.ImageModel(modelID) calls the legacy Stability v1 REST API — POST /v1/generation/{modelID}/text-to-image — so modelID must be a v1 "engine" ID. Stability's newer SD3/Stable Image v2beta models (sd3-large, sd3-medium, core, ultra, etc.) are not reachable through this provider, since they live under a different endpoint (/v2beta/stable-image/...) that this package does not implement.

Image Generation​

Model IDOutputBest For
stable-diffusion-xl-1024-v1-0PNG bytesHigh quality text-to-image generation
stable-diffusion-v1-6PNG bytesFast text-to-image generation

Provider-Specific Features​

Text-to-Image​

result, err := ai.GenerateImage(ctx, ai.GenerateImageOptions{
    Model:  model,
    Prompt: "A serene landscape",
    Size:   "1024x1024",
})

Examples​

Basic Image Generation​

package main

import (
    "context"
    "fmt"
    "log"
    "os"

    "github.com/digitallysavvy/go-ai/pkg/ai"
    "github.com/digitallysavvy/go-ai/pkg/providers/stability"
)

func main() {
    provider := stability.New(stability.Config{
        APIKey: os.Getenv("STABILITY_API_KEY"),
    })

    model, err := provider.ImageModel("stable-diffusion-xl-1024-v1-0")
    if err != nil {
        log.Fatal(err)
    }

    result, err := ai.GenerateImage(context.Background(), ai.GenerateImageOptions{
        Model:  model,
        Prompt: "A futuristic cityscape at sunset, photorealistic",
        Size:   "1024x1024",
    })
    if err != nil {
        log.Fatal(err)
    }

    // Save image
    imageData := result.Images[0].Data
    err = os.WriteFile("output.png", imageData, 0644)
    if err != nil {
        log.Fatal(err)
    }

    fmt.Println("Image saved to output.png")
}

Batch Generation​

prompts := []string{
    "A red sports car",
    "A blue sports car",
    "A green sports car",
}

for i, prompt := range prompts {
    result, err := ai.GenerateImage(ctx, ai.GenerateImageOptions{
        Model:  model,
        Prompt: prompt,
    })
    if err != nil {
        log.Printf("Failed to generate %d: %v", i, err)
        continue
    }

    filename := fmt.Sprintf("car_%d.png", i)
    os.WriteFile(filename, result.Images[0].Data, 0644)
}

Multiple Images​

n := 4
result, err := ai.GenerateImage(ctx, ai.GenerateImageOptions{
    Model:  model,
    Prompt: "A set of colorful product concept images",
    N:      &n,
    Size:   "1024x1024",
})
if err != nil {
    log.Fatal(err)
}

for i, image := range result.Images {
    filename := fmt.Sprintf("concept_%d.png", i)
    os.WriteFile(filename, image.Data, 0644)
}

Best Practices​

  1. Prompt Engineering

    • Be specific and detailed
    • Include style keywords
    • Mention quality (4k, detailed, etc)

  2. Cost Optimization

    • Start with one image before increasing N
    • Batch similar requests
    • Cache generated images

  3. Performance

    • Use appropriate model for task
    • Parallelize generations

Rate Limits & Pricing​

Rate Limits​

Varies by plan - typical limits:

  • 150 images/min (standard)
  • 500 images/min (premium)

Cost Management​

func estimateImageCost(model string, count int) float64 {
    prices := map[string]float64{
        "stable-diffusion-xl-1024-v1-0": 0.04,
        "stable-diffusion-v1-6":         0.02,
    }

    return prices[model] * float64(count)
}

See Also​