AI provider configuration

Every AI provider supported by Aimogen Pro, where each credential goes, what it unlocks, and how the plugin decides which provider handles a request.

Aimogen Pro talks to twelve textual AI providers plus any OpenAI-compatible endpoint, and to a separate set of services for images, video, speech, search and data.

Every credential lives in one place: Aimogen Pro › Settings › API Keys.

Textual AI providers#

Supporting services#

How the provider is chosen#

Aimogen Pro does not have a global "current provider" for text generation. It derives the provider from the model you select, per feature.

Each model selector in the plugin lists every model from every configured provider, annotated with its provider name. When a request runs, the plugin matches the model name against its provider tables and routes accordingly.

One setting does act globally: Settings › API Keys › Main API Service Provider Selector. It has three values and only affects the OpenAI-compatible slot:

Anthropic, Google, xAI, Perplexity, Groq, NVIDIA, OpenRouter, Hugging Face and Ollama are unaffected by that selector — their models are routed by name regardless.

Setting keys#

For developers reading settings directly, the credential lives in aiomatic_Main_Settings under these keys. All are textareas holding one key per line.

api_selector chooses which destination the app_id slot targets: openai, azure or custom.

See Data model.

Multiple keys per provider#

Every provider key field is a textarea, one key per line. This gives you:

• Higher throughput. Requests spread across keys, so per-key rate limits bite later.
• Resilience. A key that starts failing can be taken out of rotation automatically. See Smart API-key rotation.

Settings › API Keys › Use Multiple API Keys When Creating The Same Post controls whether a single post sticks to one key for all its requests, or picks a fresh key per request. Sticking to one key produces more consistent rate-limit behaviour for long articles; rotating spreads load.

Testing a key#

Every key field has a Test API key button beside it. It performs a real, minimal request against the provider and reports success or the provider error text verbatim.

Use it whenever you paste a key. An untested key is the most common cause of scheduled runs that produce nothing at all.

Where keys are stored#

Provider credentials are stored in the WordPress options table, inside the aiomatic_Main_Settings option, as plain text. This is the same storage model used by the overwhelming majority of WordPress plugins that talk to third-party APIs.

Consequences and mitigations are covered in Where credentials are stored. In short: anyone with database access or the manage_options capability can read them, so treat both as privileged.

If you would rather not store keys in the database at all, supply them from code with the per-provider filters:

add_filter( 'aimogen_openai_api_key', function () {
    return defined( 'MY_OPENAI_KEY' ) ? MY_OPENAI_KEY : '';
} );

Every provider has an equivalent filter. See Filter reference.

Rate limiting and throttling#

Two settings deliberately slow the plugin down so you stay inside provider quotas:

Delay Between API & Scraping Requests (ms). A pause inserted between consecutive text and scraping calls. Raise it if you see HTTP 429 during bulk runs.

Delay Between API Requests For AI Image Generator APIs (ms). The same idea for image endpoints, which usually have much tighter limits than text endpoints.

Both live on Settings › API Keys. For automatic handling of 429s rather than blanket slowdown, see Reliability Mode.

Provider-specific Responses API#

Several providers now offer a "Responses" API alongside the classic chat completions API. Aimogen Pro exposes a toggle per provider under Settings › Advanced AI Settings:

Security considerations#

• Never paste a key into a support ticket, a public forum or a screenshot. Rotate it immediately if you do. The plugin diagnostics deliberately never print raw keys.
• Use restricted keys where the provider supports them. Scope to the models and spend you need.
• Set provider-side spending caps. The plugin usage limits protect your site; only the provider dashboard protects your card.
• Rotate keys when staff leave, since anyone with manage_options could read them.

See Security and privacy.

Common errors, by cause#

Start with OpenAI, or jump to the provider you use.