AI Settings

Configure the AI provider, model, and automatic schema generation — using your own OpenAI, Anthropic (Claude), or Google Gemini API key.

Navigation Path

WordPress Admin  →  SchemaEngine AI  →  Settings  →  AI Settings

Sections on This Page

15.1.  General AI Settings

15.2.  API Configuration

15.3.  How to Enable Auto-Generation (Full Workflow)

15.4.  Reset & Re-run Auto-Generation

15.5.  Server & Cron Requirements

15.6.  Supported Providers & Models

①	General AI Settings Master toggle · confidence threshold · FAQ count · auto-generation

 

15.1  General AI Settings #

These settings govern how the AI behaves when generating schema automatically. They affect both the manual generation panel in the block editor (Step 14) and the background auto-generation triggered on post publish.

Field	Description
Enable AI	Master toggle — must be ON for any AI generation to work. Default: OFF. When ON, the AI generation panel in the Gutenberg editor becomes active. When OFF, all AI features are disabled site-wide regardless of other settings.
Confidence Threshold	Numeric value 0–100. If the AI’s confidence in the generated schema falls below this number, it falls back to a generic WebPage schema instead. Default: 60. Raise this value (e.g. 80) to enforce stricter quality; lower it to accept more uncertain outputs.
FAQ Count	How many FAQ question-and-answer pairs to generate when auto-generating schema for FAQPage content or when no FAQ block exists on the post. Default: 8. This setting applies to automatic generation — the Number of FAQs stepper in the block editor panel (Step 14) overrides this for manual generation.
Auto-Generate	Toggle — when ON, SchemaEngine AI automatically generates schema using AI for every eligible post based on Post Type Mapping and Site Info settings. Each publish or update triggers a background API call to your AI provider. Default: OFF. See Section 15.3 for the full activation workflow.
Reset Auto-Generation	A link that appears below Auto-Generate when it is ON: “Reset & Re-run Auto-Generation”. Clicking this clears the completion flag on all posts and queues them for fresh schema generation. See Section 15.4 for full details.

Field	Validation	Note
Enable AI	○  Optional	Default OFF. All AI features are inactive when OFF. No error displayed — features simply do not work.
Confidence Threshold	○  Optional	Default: 60. Range 0–100. Start at 60, observe output quality, then adjust.
FAQ Count	○  Optional	Default: 8. Changed from 5 in v3.0.0. Applies to auto-generation only.
Auto-Generate	○  Optional	Default OFF. Requires WP Cron (or server cron) to be running — see Section 15.5.

↑ Back to Contents

②	API Configuration Provider · API key · model selection · max tokens

 

15.2  API Configuration #

Select your AI provider and enter the credentials required to authenticate. The fields shown below the API Provider dropdown change dynamically to match the selected provider.

Field	Description
API Provider	Dropdown — choose which AI service handles schema generation. Options: OpenAI, Anthropic (Claude), Google Gemini. The screenshot shows Anthropic (Claude) selected.  Default: OpenAI.
API Key	Your secret key from the selected provider. The field shown changes label depending on the provider selected (e.g. “Anthropic API Key”, “OpenAI API Key”, “Gemini API Key”). The eye icon toggles key visibility. Stored in your WordPress database — transmitted only to your chosen AI provider.
Model	Dropdown — select the AI model for schema generation. Options are specific to the selected provider. The screenshot shows “Claude Sonnet 4.6” selected for Anthropic. Default: GPT-4.1 for OpenAI, Claude Sonnet 4.6 for Anthropic.
Max Tokens	Maximum tokens the AI can use per response. Default: 4096. Higher values produce more detailed schema but cost more per API call. Do not reduce below 2048 — very low values cause the AI to truncate the schema output. For most posts 4096 is sufficient.

Field	Validation	Note
API Provider	✖  Error (required)	A provider must be selected for AI generation to initialise.
API Key	✖  Error (required)	Required. An invalid or expired key causes a generation error.
Model	⚠  Warning (recommended)	Select explicitly for predictable output quality. Leaving blank defaults to the provider’s current default model.
Max Tokens	○  Optional	Default: 4096. Only increase above 4096 for very long or complex posts with many schema types.

↑ Back to Contents

 

15.3  How to Enable Auto-Generation (Full Workflow) #

Auto-Generate is the most powerful AI Settings feature — once properly configured, it generates schema for every eligible post automatically without any manual action. Follow these steps in order.

⚠️  Complete all prerequisites before enabling Auto-Generate Enabling Auto-Generate without completing the steps below will result in incomplete or incorrect schema generation. Take each step in sequence.

1

Configure Post Type Mapping Navigate to Settings → Schema → Post Type Mapping (Step 4). For every post type you want AI to generate schema for, ensure: (a) a Schema Type is selected — not left empty, and (b) the Auto-Generate toggle for that post type is ON. Why this matters: The AI uses the Schema Type setting to determine what type of JSON-LD to generate for each post type. If Schema Type is not set, the AI has no instruction on what markup to produce.

2	Complete Site Info Ensure Settings → Schema → Site Info (Step 2) is fully filled in — especially Name, Logo, Address, and Telephone for LocalBusiness setups. The AI uses Site Info data to populate Organisation, WebSite, and BreadcrumbList schema properties.

3	Configure AI Settings — enable and save In Settings → AI Settings: (1) Set Enable AI to ON. (2) Select your API Provider. (3) Enter your API Key. (4) Select a Model. (5) Set Auto-Generate to ON. (6) Click Save Change. Do not click Reset yet.

4	Verify WP Cron is running Auto-generation is processed via WordPress Cron (WP-Cron) in the background. Before clicking Reset, confirm cron is working on your server. See Section 15.5 for how to check and configure cron.

5

Click Reset & Re-run Auto-Generation Once everything above is confirmed, click the “Reset & Re-run Auto-Generation” link that appears below the Auto-Generate toggle. This clears any existing completion flags and queues all eligible posts for AI schema generation. The process runs in the background — see Section 15.4 for details.

💡  New posts generate automatically after this point Once Auto-Generate is ON and Reset has been run, any new post published or updated on an eligible post type will automatically receive AI-generated schema — no further manual action needed.

↑ Back to Contents

 

15.4  Reset & Re-run Auto-Generation #

The “Reset & Re-run Auto-Generation” link is a critical tool for triggering bulk AI schema generation across all existing posts. It only appears when Auto-Generate is set to ON.

When to use Reset & Re-run

Scenario	Action recommended
First-time setup — generate schema for all existing posts	Click Reset after fully configuring AI Settings and verifying cron is running
Switched AI provider or model	Click Reset to regenerate all post schemas using the new provider/model
Updated Site Info data significantly	Click Reset to regenerate schemas that reference Site Info (Organisation, WebSite, etc.)
Schema Type Mapping changed for a post type	Click Reset to regenerate schema under the new type for all affected posts
Some posts missed during initial generation	Click Reset to catch any posts the cron may have skipped
Schema looks stale or incorrect site-wide	Click Reset to force a full re-generation pass

⚠️  Reset triggers one API call per post — check your usage limits Clicking Reset on a site with 500 posts will queue 500 API calls to your AI provider. Each call incurs usage costs. On large sites, monitor your provider’s usage dashboard after clicking Reset. Consider running Reset during off-peak hours to spread the load.

↑ Back to Contents

 

15.5  Server & Cron Requirements #

Auto-Generate relies on WordPress Cron (WP-Cron) to process schema generation in the background. If WP-Cron is not running correctly, auto-generation will not work — even with all settings configured correctly.

How to verify WP-Cron is working #

1. Install and activate the WP Crontrol plugin (free from wordpress.org/plugins/wp-crontrol).
2. Go to Tools → Cron Events in the WordPress admin.
3. Look for SchemaEngine AI cron events in the list — they should be scheduled.
4. Check that events are firing on schedule and not stuck with overdue timestamps.

Recommended: Use a Real Server Cron Job #

For reliable Auto-Generation on any site — especially low-traffic or high-post-count sites — disable WP-Cron and replace it with a real server-level cron job. This is the most reliable configuration.

🖥️  Setting Up a Real Server Cron Job Step 1 — Disable WP-Cron. Add this line to your wp-config.php file (before the line that says “That’s all, stop editing”): define(‘DISABLE_WP_CRON’, true); Step 2 — Add a server cron job. Log in to your server via SSH or cPanel → Cron Jobs, and add a cron job that runs every minute: */1 * * * * wget -q -O – https://yourdomain.com/wp-cron.php?doing_wp_cron >/dev/null 2>&1 Replace yourdomain.com with your actual domain. This ensures cron tasks run every minute reliably, regardless of site traffic.

PHP Execution Time

Each AI API call may take 5–15 seconds depending on post length and the AI provider. Ensure your server’s PHP max_execution_time is set to at least 120 seconds to prevent schema generation from timing out mid-process. Check this in your php.ini or ask your hosting provider.

Server requirement	Recommended value
PHP max_execution_time	120 seconds or higher (default is often 30–60s)
PHP memory_limit	256M or higher
WP-Cron or server cron	Must be running every minute for reliable generation
HTTPS	Required — AI provider APIs reject plain HTTP connections
Outbound HTTP requests	Server must allow outbound HTTPS connections to api.openai.com, api.anthropic.com, or generativelanguage.googleapis.com

↑ Back to Contents

 

15.6  Supported Providers & Models #

Three AI providers are supported. Each requires a separate account and API key. You can switch providers at any time — simply select the new provider, enter its API key, and save.

Model Comparison #

Provider	Model	Schema Quality	Speed	Best For
OpenAI	gpt-4.1	★★★★★	Fast	Best overall — recommended starting point
OpenAI	gpt-4o	★★★★★	Fast	Multimodal — handles image-rich posts
OpenAI	gpt-4o-mini	★★★★	Fastest	Budget option — simple posts
OpenAI	gpt-3.5-turbo	★★★	Fastest	Lowest cost — high volume bulk generation
Anthropic	claude-sonnet-4-6	★★★★★	Fast	Shown in screenshot — highly recommended
Anthropic	claude-haiku-4	★★★★	Fastest	Fast and cost-effective
Anthropic	claude-opus-4	★★★★★	Moderate	Complex / long-form content
Gemini	gemini-2.5-pro	★★★★★	Moderate	Best Gemini model
Gemini	gemini-2.5-flash	★★★★	Fast	Speed + quality balance
Gemini	gemini-2.0-flash	★★★	Fastest	Free tier — ideal for testing

💡  Security — API key best practices Create a dedicated API key specifically for SchemaEngine AI at your provider dashboard. This lets you revoke it independently if needed. Set usage limits or spending caps at your provider to prevent unexpected billing if the key is ever compromised. Never include your API key in publicly accessible files or version control.

💾  Save Changes
After updating any field on this page, click the Save Change button in the bottom-right corner. Both General AI Settings and API Configuration are saved together in a single action. Always save before clicking Reset & Re-run Auto-Generation.

↑ Back to Contents