Image and video providers
Every image and video generation service Aimogen Pro supports, where each credential goes, the models and parameters available, and how to choose between them.
Aimogen Pro can generate images through seven providers and video through one. All credentials live in Settings › API Keys; behaviour defaults live in Settings › AI Images and Settings › AI Videos.
This page covers the credentials and provider-specific options. For how images are used inside content, see Image generation and Featured images and in-post media.
At a glance#
| Provider | Credential field | Cost model | Notable strength |
|---|---|---|---|
| OpenAI | shares the main OpenAI key | Per image | Prompt adherence, text rendering |
| Stability.AI | Stability.AI API Keys | Credits | Fine-grained control, image editing and upscaling |
| shares the Google AI Studio key | Per image | Imagen quality, photographic realism | |
| Ideogram | Ideogram API Keys | Subscription plus credits | Text inside images, typography |
| Replicate | Replicate API Keys | Per second of compute | Model variety including Flux |
| GoAPI (Midjourney) | GoAPI.AI (Midjourney) API Keys | Subscription | Midjourney aesthetics |
| Cloudflare Workers | Cloudflare Worker URL and Cloudflare Worker Secret | Free tier on Cloudflare | No per-image cost |
| OpenAI Sora | shares the main OpenAI key | Per video | Video generation |
OpenAI images#
Uses the key already configured for OpenAI.
Models: gpt-image-2 (default), gpt-image-1.5, gpt-image-1, gpt-image-1-mini,
chatgpt-image-latest.
Configure the default under Settings › AI Images: AI Image Model and AI Generated Image Size.
Stability.AI#
Key: beta.dreamstudio.ai membership page → Settings › API Keys › Stability.AI API Keys.
Models (Image Model under Settings › AI Images):
stable-diffusion-ultra, stable-diffusion-core, stable-diffusion-3-0-large,
stable-diffusion-3-0-turbo, stable-diffusion-3-0-medium, sd3.5-large, sd3.5-large-turbo
(the plugin default), sd3.5-medium, plus the legacy stable-diffusion-xl-1024-v1-0 and
stable-diffusion-v1-6.
Provider-specific parameters, all on Settings › AI Images:
| Setting | What it does |
|---|---|
Steps | Diffusion steps. More steps means more detail and more cost. Diminishing returns past roughly 30–50 |
CFG Scale | How strictly the image follows the prompt. Low values are more creative, high values more literal and sometimes over-saturated |
Seed | Fix it to reproduce the same image from the same prompt. Leave empty for variety |
Clip Guidance Preset | Guidance strategy for the older SD models |
Style Preset | Applies a named aesthetic, for example photographic or digital art |
Sampler | Sampling algorithm for the older SD models |
Stability is also the backend for image editing and upscaling — background removal, search-and-replace, search-and-recolour, inpaint, outpaint, erase and background replacement, all added in 2.8.8.
Google (Imagen and Gemini images)#
Uses the Google AI Studio key.
Models: imagen-4.0-generate-001 (default), gemini-3.1-flash-image-preview,
gemini-3-pro-image-preview, gemini-2.5-pro, gemini-2.5-flash, gemini-2.5-flash-lite.
Set with Settings › AI Images › Google Image Model.
Ideogram#
Key: ideogram.ai/manage-api → Settings › API Keys › Ideogram API Keys.
Models: V_3, V_2_TURBO, V_2, V_1_TURBO, V_1.
Provider-specific parameters:
| Setting | What it does |
|---|---|
Magic Prompt | Lets Ideogram expand your prompt automatically. Better results from short prompts; less control |
Aspect Ratio | Output proportions |
Style | Named style preset |
Color Palette | Constrains the palette |
Ideogram is the strongest of these at rendering legible text inside an image, which matters for banners and thumbnails with words on them.
Replicate#
Key: replicate.com/account/api-tokens → Settings › API Keys › Replicate API Keys.
Replicate hosts many models, so you choose one. The plugin caches the available list in the
aiomatic_replicate_model_list option and ships a default API version constant for the fallback model.
Provider-specific parameters:
| Setting | What it does |
|---|---|
Replicate Image Model | Which hosted model to run |
Prompt Strength | How strongly the prompt overrides an input image, for image-to-image |
Aspect Ratio | Output proportions |
Number Of Inference Steps | Quality versus cost and time |
Scheduler | Sampling scheduler |
Custom Parameters | Free-form JSON passed to the model, for parameters the plugin has no field for |
Billing is per second of compute, so a slow model with many steps costs meaningfully more than a fast one.
GoAPI (Midjourney)#
Midjourney has no official public API. Aimogen Pro reaches it through GoAPI.
Key: Settings › API Keys › GoAPI.AI (Midjourney) API Keys.
Settings: Midjourney Image Engine and Midjourney Image Model under Settings › AI Images.
Midjourney generations are asynchronous and slower than the other providers. Expect tens of seconds.
Cloudflare Workers#
Added in 2.8.0. You deploy a small Worker on your own Cloudflare account that fronts Cloudflare Workers AI image models. Because Cloudflare has a generous free tier, this is the only option here with no per-image cost for typical volumes.
Settings › API Keys
| Setting | Value |
|---|---|
Cloudflare Worker URL | The URL of your deployed Worker, one per line |
Cloudflare Worker Secret | The shared secret your Worker checks, one per line |
The Knowledge Base carries the step-by-step Worker deployment guide, linked directly from the settings help text: How to create unlimited AI images for free on WordPress using Cloudflare Workers.
Quality is below the paid providers, but for decorative or section images it is often good enough.
OpenAI Sora (video)#
Uses the OpenAI key.
Models: sora-2 (default) and sora-2-pro. Resolution and duration options are defined per model in
the plugin. Configure under Settings › AI Videos.
See Video generation.
Shared image settings#
These apply regardless of provider, under Settings › AI Images:
| Setting | What it does |
|---|---|
Copy Royalty Free / AI Images From Post Content | Whether generated images are sideloaded to your Media Library, or offloaded to Amazon S3, Cloudflare R2, DigitalOcean Spaces, Wasabi or a generic S3 endpoint. Disabled means images stay hotlinked at the provider URL, which will break when the provider expires them |
AI Generated Image Resize Width / Height | Resize on import. Only applies when images are copied locally |
Resize Quality | JPEG quality for the resize |
Disable Compression / Compression Quality | Control recompression on import |
Delay Between API Requests For AI Image Generator APIs (ms) | Throttle, on Settings › API Keys. Image endpoints rate-limit far more aggressively than text endpoints |
See Cloud storage for the offload targets.
Choosing a provider#
| Goal | Suggestion |
|---|---|
| Best prompt adherence | OpenAI gpt-image-2 |
| Text inside the image | Ideogram |
| Photographic realism | Google Imagen |
| Fine control, editing, upscaling | Stability.AI |
| Distinct artistic style | Midjourney via GoAPI |
| No per-image cost | Cloudflare Workers |
| A specific open model such as Flux | Replicate |
Common problems#
Images generate but do not appear in posts
Copying is disabled and the provider URL expired. Turn on
Copy Royalty Free / AI Images From Post Content.
429 during bulk runs
Image endpoints have tight limits. Raise
Delay Between API Requests For AI Image Generator APIs (ms).
Prompt rejected Every provider applies content policy. Rephrase.
Midjourney requests time out
Generation is asynchronous and slow. Raise PHP max_execution_time.
Related#
Still stuck? Open a support ticket and include the diagnostics from Aimogen Pro › System & Logs › System Info.