Image Generation

Use these models to generate whatever images you can (or can't!) imagine.

Image Generation Models on Sentry Block

Sentry Block supports a range of powerful diffusion models for AI-generated imagery. Here’s a quick overview of available models and their supported parameters:

Parameter

StableDiffusion XL

Flux.1 schnell

prompt

negative_prompt

width

height

num_steps

guidance_scale

seed

Both models are capable of producing high-quality images, each with its own visual style. We recommend trying both to see which best fits your project’s needs.


Getting Started: Using the Models

Through the Sentry Block UI

  1. Visit the Sentry Block platform.

  2. Log into your account and confirm you have an active credit balance.

  3. Navigate to “Serverless Endpoints” and choose your image generation model.

  4. Set your parameters, input your prompt, and hit Enter to generate your image.


Key Parameters

  • Prompt (required) – Text prompt guiding image creation.

  • Negative Prompt – Text that defines what to avoid in the output.

  • Width / Height – Dimensions of the output image.

  • Steps (num_steps) – Controls the number of inference steps (higher = more refined output).

  • Guidance Scale – Strength of adherence to prompt (higher = more literal, but less artistic).

  • Seed – Optional. Use this to make results reproducible.

Bolded parameters are universal. Others may vary depending on the model selected.


Using the API Directly

You can generate images programmatically via the Sentry Block API. Just include your API key and model of choice. Below are some examples.

⚠️ Don't forget to check the [API Reference] for authentication and error handling tips.


cURL Example


Python Example


JavaScript Example


Model Mapping

Use the following names to specify your model in API requests:

  • StableDiffusion XL 1.0stabilityai/stable-diffusion-xl-base-1.0

  • Flux.1 schnellblack-forest-labs/FLUX.1-schnell


Response Example

A successful image generation request will return a JSON response that looks like this:

To view the image, decode the b64_json value using any base64 image decoder.

Last updated