WordPress 7.0 Connectors Integration – Aimogen Pro Documentation - Aimogen

Last updated: June 2026
Applies to: Aimogen Pro (Aiomatic) with WordPress 7.0+
Connectors page: Settings → Connectors (/wp-admin/options-connectors.php)

Overview

WordPress 7.0 introduces a centralized Connectors screen where API keys and credentials for AI providers are stored once and shared across plugins. By default, WordPress expects separate “AI Provider” plugins for each service (OpenAI, Anthropic, Google and so on).

Aimogen Pro’s Connectors integration lets you use the same API keys, models and providers you already configured in Aimogen Pro to power native WordPress AI features – excerpt generation, AI blocks, the AI plugin and anything built on the WordPress AI Client – without installing a separate provider plugin for every service.

With one click, Aimogen Pro can:

Register all 10 supported providers on the Connectors page
Keep credentials in sync between Aimogen Pro and WordPress
Supply built-in AI Client adapters for providers WordPress does not ship plugins for (Groq, NVIDIA, xAI, OpenRouter, Perplexity, Hugging Face, Ollama and more)
Let you pick which text and image models WordPress AI should prefer

What You Need

Requirement	Details
WordPress	7.0 or later
Aimogen Pro	Active, with the Connectors module loaded
WordPress AI	The official AI plugin (recommended for editor AI features)
PHP	Same as Aimogen Pro requirements
Permissions	Administrator (`manage_options`) to enable the integration

The Connectors page itself requires WordPress core’s built Connectors UI. If the page shows a build error, update WordPress to a complete 7.0 release.

Supported Providers

When Aimogen Pro manages connectors, these providers appear on the Connectors page as “{Provider} via Aimogen Pro”:

Provider	Aimogen Pro setting	Notes
OpenAI	OpenAI API Key (`app_id`)	GPT and related models
Anthropic	Claude API Key (`app_id_claude`)	Claude models via OpenAI-compatible API
Google	Google API Key (`app_id_google`)	Gemini / Imagen via Google’s OpenAI-compat endpoint
OpenRouter	OpenRouter API Key (`app_id_openrouter`)	Large multi-provider catalog
Perplexity	Perplexity API Key (`app_id_perplexity`)	Sonar and related models
Groq	Groq API Key (`app_id_groq`)	Fast inference
NVIDIA	NVIDIA API Key (`app_id_nvidia`)	NIM hosted models
xAI	xAI API Key (`app_id_xai`)	Grok models
Hugging Face	Hugging Face Token (`app_id_huggingface`)	Inference router models
Ollama	Ollama Server URL (`ollama_url`)	Local/self-hosted models; URL is treated as the connector credential

A provider only shows as Connected when:

The corresponding key or URL is set in Aimogen Pro → Settings → API Settings (tab 2) and
The provider is reachable and correctly configured in the WordPress AI Client registry.

Providers without a configured credential show Set up and redirect you to Aimogen Pro settings – not the core inline key dialog.

Quick Start

Step 1 – Configure keys in Aimogen Pro

Go to Aimogen Pro → Settings.
Open the API Settings tab (#tab-2).
Enter API keys (and your Ollama server URL, if used).
Save settings.

This is the single source of truth for credentials when Aimogen Pro is managing connectors.

Step 2 – Open the Connectors page

Go to Settings → Connectors in WordPress admin.

You will see an Aimogen Pro banner at the top of the page.

Step 3 – Enable managed mode

Click Enable (or Take over if another plugin is already managing connectors).

The page reloads. Connector cards for your configured providers should appear as connected. The banner turns green and shows Managed by Aimogen Pro.

Step 4 – Choose WordPress AI models (optional)

In the green banner, use the Text and Image dropdowns to choose which provider and model WordPress AI features should prefer:

Automatic (WordPress default) – WordPress picks from its own preference list.
Specific model – Your choice is saved and tried first for excerpt generation, AI blocks, etc.

Click the refresh icon to reload the model list from connected providers.

Operating Modes

Aimogen Pro uses three internal modes:

Mode	Behavior
Observe (default)	Banner is shown; Aimogen Pro does not take over connectors. WordPress behaves normally.
Managed	Aimogen Pro registers all providers, bridges keys, supplies AI Client adapters, handles model preferences and powers WordPress AI through your Aimogen configuration.
Off	Integration disabled entirely (no banner, no registry changes).

You can return to Observe anytime with the Stop button in the banner. Your Aimogen Pro keys are not deleted.

Dismiss hides the banner in Observe mode only. If another plugin conflicts while you are in Managed mode, the banner stays visible until resolved.

Managing API Keys

Where to edit keys

When Aimogen Pro is in Managed mode:

Always edit credentials in Aimogen Pro → Settings → API Settings (tab 2).
Clicking Edit or Set up on any connector card redirects there automatically.

This is intentional. Editing keys through WordPress’s native connector dialog can cause validation issues for providers that do not have official WordPress provider plugins.

Two-way sync

Aimogen Pro keeps credentials synchronized:

Direction	When it happens
Aimogen Pro → Connectors	When you save Aimogen settings, on admin load (`initial_sync`) and when you enable Managed mode
Connectors → Aimogen Pro	When a connector option is updated through the REST API (if you change a key outside Aimogen)

Safety rules:

Empty connector values never wipe existing Aimogen keys (protects against REST validation clearing options).
Multi-line keys (round-robin keys in Aimogen) sync using the first line to Connectors; additional lines are preserved in Aimogen.
Ollama syncs the server URL into the connectors_ai_ollama_api_key option.

WordPress AI Model Selection

WordPress AI features (excerpt generation, AI-powered blocks, etc.) choose models from a preference list. Aimogen Pro hooks into that list when Managed mode is active.

How selection works

Open Settings → Connectors with Managed mode enabled.
In the banner, pick a Text model and/or Image model.
Your choice is saved in WordPress options.
Aimogen Pro prepends your selection to WordPress’s wpai_preferred_text_models and wpai_preferred_image_models filters so it is tried first.

Which models appear in the picker

Only providers that are actually connected (credential present + AI Client reports configured) appear in the dropdown. Models come from Aimogen’s curated lists and, where available, live provider catalogs (OpenRouter, Groq, etc.).

If no providers are connected, the picker shows: “Connect a provider above to choose models.”

Automatic mode

Selecting Automatic (WordPress default) clears your saved preference for that type (text or image) and lets WordPress use its built-in model ordering.

How It Works (Plain Language)

┌─────────────────────┐ ┌──────────────────────┐ ┌─────────────────────┐

│ Aimogen Pro │ │ WordPress │ │ AI Provider APIs │

│ API Settings │────▶│ Connectors page │────▶│ (OpenAI, Groq, │

│ (keys + Ollama URL)│ sync│ + AI Client │ │ Anthropic, …) │

└─────────────────────┘ └──────────────────────┘ └─────────────────────┘

│

▼

WordPress AI features

(excerpt, blocks, AI plugin)

Registry – On wp_connectors_init, Aimogen Pro registers one connector card per provider.
Adapters – Built-in AI Client provider classes route requests to each API (OpenAI-compatible endpoints where applicable).
Credentials gate – WordPress AI’s “do I have valid AI credentials?” checks are satisfied when Aimogen has configured providers.
Caching – Model catalogs are cached in WordPress transients and warmed in the background so editor features stay fast and reliable.
Approval – The WordPress AI plugin’s Connector Approval experiment is auto-approved for ai/ai.php and Aimogen Pro when Managed mode is active, so outbound AI requests are not blocked.

Official provider plugins (e.g. “AI Provider for OpenAI”) are never overridden if already installed – Aimogen’s adapters only fill gaps.

Built-in Provider Adapters

WordPress normally requires a separate plugin per provider. Aimogen Pro includes adapters for all 10 providers so you do not need:

AI Provider for Groq
AI Provider for OpenRouter
AI Provider for Ollama
…and so on

Each adapter:

Uses the provider’s OpenAI-compatible API where available
Lists models from live /models endpoints when possible
Falls back to Aimogen’s curated model lists when an endpoint is unavailable (Perplexity, Anthropic static list)
Respects Anthropic advanced model rules (see below)

If you install an official WordPress provider plugin for a service, that plugin takes precedence.

Ollama (Local Models)

Ollama is supported as a first-class connector:

Credential: Your Ollama server URL from Aimogen Pro (e.g. http://localhost:11434)
Connected when: The URL is set and Ollama responds to model listing
Set up / Edit: Redirects to Aimogen Pro API settings

Tips:

Use a reachable URL from your WordPress server (not localhost if WordPress runs in Docker and Ollama does not).
Aimogen supports multiple Ollama URLs (one per line) for round-robin; the first line is used for WordPress AI.
If Ollama is temporarily unreachable, a cooldown prevents the Connectors page from hanging while loading models.

Anthropic Advanced Models

Newer Claude models (Opus 4.6+, Sonnet 4.6, Fable 5, Mythos 5, etc.) reject classic sampling parameters such as temperature and top_p when called through OpenAI-compatible APIs.

Aimogen Pro automatically strips those parameters for affected models so WordPress AI features do not fail with:

400 - temperature is deprecated for this model

No manual configuration is required.

Conflicts With Other Plugins

Some plugins also manage WordPress 7 Connectors:

Plugin	Detected automatically
AI Power / AI Puffer	Yes
AI Engine	Yes
Other plugins	Detected if connector names show “via {Plugin Name}”

If a conflict is detected

The banner turns amber and explains which plugin is managing connectors.

Your situation	Action
Another plugin manages connectors; you want Aimogen	Click Take over – the other plugin is switched to Observe mode (its keys are untouched)
Aimogen manages connectors; another plugin overrode you	Click Take back control
You prefer the other plugin	Click Stop in Aimogen’s banner to return to Observe mode

Only one plugin should be in Managed mode at a time.

Connector Approval (WordPress AI Plugin)

The WordPress AI plugin includes a Connector Approval security experiment. It can block outbound AI requests unless a plugin is explicitly approved to use a connector’s credentials.

When Aimogen Pro is in Managed mode, it automatically approves:

The WordPress AI plugin (ai/ai.php)
Aimogen Pro itself

for all Aimogen-managed providers. Other plugins still go through the normal approval flow.

If you see errors mentioning a connector “has not been approved”, reload the Connectors page after enabling Managed mode, or disable Managed mode and re-enable it to re-run approval sync.

Troubleshooting

Connectors page is blank or uses high CPU

This was addressed in Aimogen Pro by avoiding DOM patterns that conflict with WordPress’s React Connectors UI. Update Aimogen Pro to the latest version. If the issue persists, hard-refresh the page (Ctrl+F5).

Provider shows “Set up” but I entered a key in Aimogen Pro

Confirm the key is saved in Aimogen Pro → Settings → API Settings.
Ensure Managed mode is enabled.
Reload the Connectors page (Aimogen syncs on admin load).
For Ollama, confirm the server URL field is filled, not an API key field.

`unsupported_model` in the editor

Usually means the model catalog could not be loaded. Aimogen Pro caches catalogs in transients and warms them in the background. Try:

Reload the Connectors page.
Click the model picker refresh icon.
Confirm the provider is connected and the chosen model exists for that provider.

`prompt_network_error` or connector not approved

Often the Connector Approval guard blocking requests before Managed approval ran. Re-enable Managed mode or check that the WordPress AI plugin is active.

Wrong Google key attributed / OpenAI request mentions Google

Ensure each provider’s key in Aimogen Pro is in the correct field. The same key value in both OpenAI and Google fields can confuse connector detection. Use distinct keys per provider.

Models from disconnected providers still listed

The model picker only lists connected providers. Clear your browser cache, click refresh in the banner, or temporarily disable and re-enable Managed mode to invalidate cached model lists.

Anthropic 400 temperature error

Update Aimogen Pro. Advanced Claude models require sampling params to be omitted; this is handled automatically in current versions.

Frequently Asked Questions

Do I still need the WordPress AI plugin?
Yes, for most native WordPress AI editor features. Aimogen Pro supplies credentials and routing; the AI plugin provides the features themselves.

Do I need separate “AI Provider for …” plugins?
No, not for the 10 providers Aimogen supports – unless you prefer the official WordPress provider plugin for a specific service (OpenAI, etc.). Official plugins take precedence when installed.

Will enabling this change how Aimogen Pro generates content?
No. Aimogen’s own generation, bulk tools and workflows are unchanged. This integration adds a bridge so WordPress core AI features can use the same keys and models.

Can I use different models for Aimogen vs WordPress AI?
Yes. Aimogen Pro uses its own model settings for plugin features. The Connectors banner controls WordPress AI text/image preferences only.

What happens if I uninstall Aimogen Pro?
Connector registrations and bridged options may remain in WordPress until cleaned up. Back up keys before uninstalling; they live in both aiomatic_Main_Settings and connectors_ai_* options while Managed mode was active.

Is my API key sent to Aimogen servers?
No. Keys stay in your WordPress database and go directly from your site to the AI provider APIs, same as normal Aimogen Pro operation.

For Developers

Aimogen Pro exposes WordPress filters for customization:

Filter	Purpose
`aiomatic_wp_ai_connectors_learn_more_url`	Documentation URL in the banner (default: this page)
`aiomatic_wp_ai_connectors_known_managers`	Add competing connector manager plugins
`aiomatic_wp_ai_connectors_demote_competitors`	Set to `false` to skip demoting other managers on takeover
`aiomatic_wp_ai_connectors_approved_callers`	Customize Connector Approval auto-approve list
`aiomatic_wpai_provider_adapters`	Modify built-in AI Client adapter configuration
`aiomatic_wpai_no_sampling_model_prefixes`	Add model ID prefixes that reject temperature/top_p
`aiomatic_wp_ai_connectors_approved_callers`	Control which plugins are pre-approved for connector use

Key WordPress options:

Option	Purpose
`aiomatic_wp_ai_connectors_mode`	`observe`, `managed`, or `off`
`aiomatic_wp_ai_connectors_banner_dismissed`	Banner dismissed in Observe mode
`aiomatic_wp_ai_connectors_models`	Saved text/image model preferences
`connectors_ai_{provider}_api_key`	Mirrored connector credentials

AJAX actions (admin, nonce-protected):

aiomatic_wp_ai_connectors_set_mode
aiomatic_wp_ai_connectors_get_models
aiomatic_wp_ai_connectors_set_model

Summary

Aimogen Pro’s WordPress 7.0 Connectors integration turns your existing Aimogen API configuration into a complete AI backbone for WordPress – one settings screen, ten providers, model selection for native AI features and no maze of separate provider plugins.

Recommended workflow:

Configure all keys in Aimogen Pro → API Settings.
Enable Managed mode on Settings → Connectors.
Pick Text and Image models in the banner (or leave Automatic).
Use WordPress AI features in the editor – powered by Aimogen.

Documentation for Aimogen Pro. For support, visit Aimogen Support.