Skip to content

FAQ

SlideMind is a WordPress plugin that connects e-learning modules to AI language models (GPT, Claude, Gemini, Llama, and 200+ others). It works as a proxy: the learner interacts with AI inside the e-learning module, while all AI communication goes through your WordPress server. Your API key never reaches the browser.

SlideMind is built for:

  • Instructional designers who build e-learning courses and want to add AI interactions (tutoring, quizzes, feedback, Socratic dialogue) without writing backend code. Persistent learner memory carries context across slides, modules, and courses.
  • Training organizations that use WordPress as their LMS platform and want to enhance courses with AI while keeping costs and data under control. Learner memory lets the AI build on what a learner said or did in a previous module, so you can create adaptive learning paths.
  • Freelance e-learning developers who build courses for clients and need a reusable AI integration that works across multiple authoring tools.

No. The Wizard walks you through setup in 4 steps. The Code Generator produces ready-to-paste JavaScript for your authoring tool — you copy and paste it. No programming knowledge required.

For advanced use cases (structured JSON responses, custom variable mapping, learner memory), some familiarity with e-learning variables (e.g., Storyline variables) helps but is not required. The documentation covers each step.

Which e-learning authoring tools are supported?

Section titled “Which e-learning authoring tools are supported?”
Authoring toolStatusIntegration method
Articulate StorylineStableExecute JavaScript trigger
Universal HTML WidgetStableEmbed as iframe or web object
Adobe CaptivateBetaAdvanced Actions + Execute JavaScript
Lectora (ELB Learning)BetaRun JavaScript action
DominKnow | ClaroBetaHTML Widget
iSpring SuiteBetaWeb Object + Execute JavaScript

Beta means the code generator produces working integration code, but it has not been extensively tested in production across all versions. Community feedback is welcome.

Yes. SlideMind is a WordPress plugin — it does not require a specific LMS. It works alongside any setup where learners access e-learning content from a browser. If you use LearnDash, Tutor LMS, LifterLMS, BuddyBoss, or MemberPress, SlideMind can automatically detect learner profile fields from those plugins to personalize AI responses.

Can I use SlideMind with Moodle, Canvas, or other non-WordPress LMS?

Section titled “Can I use SlideMind with Moodle, Canvas, or other non-WordPress LMS?”

SlideMind requires WordPress. If your SCORM/xAPI packages are hosted on a WordPress site (or learners access the content through WordPress), SlideMind can be the AI backend regardless of where the e-learning module was originally authored.

If your LMS is entirely separate from WordPress, the plugin is not the right fit today. A SaaS version of SlideMind is in development and will provide the same capabilities without requiring WordPress.

Yes. When an e-learning module runs inside a SCORM player, the generated JavaScript automatically detects the SCORM API (both SCORM 1.2 and SCORM 2004) to identify the learner. This allows learner memory to persist across sessions. The HTML Widget also supports SCORM auto-detection.

SlideMind supports five LLM providers:

ProviderModelsNotes
OpenRouter200+ (Claude, GPT, Gemini, Llama, Mixtral…)Recommended — single key for all models
OpenAIGPT-4o, GPT-4, GPT-3.5…Direct OpenAI account required
AnthropicClaude Opus, Sonnet, Haiku…Direct Anthropic account required
Google (Gemini)Gemini 2.0 Flash, Pro…Google AI Studio key required
MistralMistral Large, Medium, Small…Mistral API key required

Tip: OpenRouter is the easiest way to get started — one API key gives you access to models from all major providers. You can switch models per slide without changing your key.

How much does AI cost per learner interaction?

Section titled “How much does AI cost per learner interaction?”

SlideMind itself does not charge per interaction. You pay your AI provider (e.g., OpenRouter) based on token usage. Costs vary by model:

Model tierApproximate cost per interaction
Small models (Haiku, GPT-4o Mini, Flash)$0.001—$0.005
Mid-range models (Sonnet, GPT-4o)$0.01—$0.05
Large models (Opus, GPT-4)$0.05—$0.15

These are rough estimates for a typical Q&A exchange (a few hundred tokens). Multi-turn chat sessions use more tokens. The Sandbox shows exact token counts and estimated costs for each test interaction, and the Analytics dashboard tracks cumulative usage.

Yes. Several mechanisms help:

  • Model selection per slide — use a smaller, cheaper model for simple Q&A and reserve larger models for complex tutoring.
  • Turn limiter — in Chat mode, set a maximum number of conversation turns per session.
  • API key per course — assign separate API keys to different courses for cost isolation and tracking.
  • Analytics dashboard — monitor token usage and estimated costs in real time.
  • Rate limiting — prevent abuse with per-IP request limits (configurable).

What is the difference between Free, Personal, Pro, and Agency plans?

Section titled “What is the difference between Free, Personal, Pro, and Agency plans?”
FeatureFreePersonalProAgency
Courses / Modules / Slides1 / 3 / 3UnlimitedUnlimitedUnlimited
Sites1115
Widget modesQ&A, ChatQ&A, ChatQ&A, Chat, CaptureQ&A, Chat, Capture
Custom templatesYesYesYes
Structured JSON responsesYesYesYes
Import / ExportYesYesYes
AI prompt generationYesYesYes
Learner memoryYesYes
Profile field detectionYes (9 sources)Yes (9 sources)
Analytics & learner managementYesYes
API key per courseYes
Priority support12h response

All plans include unrestricted AI model choice — you bring your own API key.


I activated the plugin but I see a warning about OpenSSL

Section titled “I activated the plugin but I see a warning about OpenSSL”

SlideMind requires the PHP OpenSSL extension to encrypt API keys at rest using AES-256-CBC. Most hosting providers have it enabled by default. If you see this warning:

  1. Contact your hosting provider and ask them to enable the openssl PHP extension.
  2. Once enabled, deactivate and reactivate SlideMind.

This is a recommendation, not an error. SlideMind works without it, but for stronger security, add a custom encryption key to your wp-config.php:

define('SM_ENCRYPTION_KEY', 'your-random-64-character-string-here');

Generate a random string using a password generator. This key encrypts your API keys at rest — if it changes, you will need to re-enter your API keys.

My API key test fails with “Invalid API key”

Section titled “My API key test fails with “Invalid API key””
  • Verify you copied the full key with no extra spaces.
  • Make sure the key matches the selected provider (e.g., an OpenRouter key starts with sk-or-).
  • Check that your API account is active and has credits.
  • If your hosting provider blocks outbound HTTPS requests, contact them to whitelist the provider’s domain (e.g., api.openrouter.ai).

My API key test fails with “Connection error”

Section titled “My API key test fails with “Connection error””

Your WordPress server cannot reach the AI provider. Common causes:

  • Hosting firewall blocks outbound HTTP/HTTPS requests. Ask your host to allow connections to the provider’s API domain.
  • wp_remote_post disabled — some shared hosts disable WordPress HTTP functions. Contact your host.
  • Timeout — the test uses a 10-second timeout. On slow connections, this may not be enough. Try again or test from a different network.

Go to SlideMind > Settings, change the Provider dropdown, enter the new API key, and click Test Connection. Existing slides will use the new provider automatically unless they have a course-level API key override.


The AI response contains markdown formatting (bold, headings, bullet points)

Section titled “The AI response contains markdown formatting (bold, headings, bullet points)”

LLMs often return markdown-formatted text (**bold**, # headings, - lists). E-learning authoring tools like Storyline or Captivate display raw text, so learners see the formatting symbols instead of formatted text.

SlideMind strips markdown from responses by default. Two toggles in SlideMind > Settings > Advanced > Response formatting control this:

  • Strip markdown (Q&A) — for responses sent to e-learning editors (enabled by default)
  • Strip markdown (Chat) — for responses in the chat widget (enabled by default)

If you still see markdown in responses, verify these toggles are enabled. If you want markdown (e.g., because you render HTML in the widget), disable the corresponding toggle.

The AI gives irrelevant or off-topic responses

Section titled “The AI gives irrelevant or off-topic responses”
  • Review your system prompt — a vague prompt leads to vague answers. Be specific about the role, topic, and constraints.
  • Use the AI Prompt Generator — on the slide edit page, click the AI prompt generation button to have the AI draft a structured system prompt based on your pedagogical goals. Refine it from there.
  • Test in the Sandbox — iterate on your prompt and see results immediately before deploying.

Add explicit length instructions in your system prompt. For example:

  • “Keep your responses under 3 sentences.”
  • “Provide a detailed explanation of at least 200 words.”

You can also pick a model that naturally produces shorter or longer responses.

The AI does not use the learner’s name or profile data

Section titled “The AI does not use the learner’s name or profile data”
  • Verify that Profile Variables are enabled for the relevant plugins in SlideMind > Settings > Profile Fields.
  • In your system prompt, reference profile variables using the @mention syntax (e.g., @first_name). The variable picker shows all available fields.
  • Profile variables only work when the learner is a logged-in WordPress user. For anonymous learners, profile variables are replaced with [field name : not available] and the AI is instructed to adapt naturally without mentioning the missing data.

Structured JSON responses are not being parsed by my e-learning module

Section titled “Structured JSON responses are not being parsed by my e-learning module”
  • In the Sandbox, verify that the AI returns valid JSON by checking the raw response.
  • Make sure your output variable names match exactly between the SlideMind configuration and your e-learning editor variables (case-sensitive).
  • Some AI models are less reliable at producing structured JSON. Try a more capable model (e.g., GPT-4o or Claude Sonnet).

SlideMind returns an error message to the learner’s interface. The default message varies by error type:

HTTP statusMessage shown
401Invalid or expired API key
402Insufficient credit on the provider account
429Rate limit exceeded, please try again later
500 / 502 / 503Server error, please try again later
Network failureConnection error

These errors are also logged in the Analytics page.


Learner memory stores variables (text, numbers, booleans) that persist across slides, modules, and courses. For example, you can capture a learner’s preferred learning style in one module and use it to personalize AI responses in another.

  • The learner must be a logged-in WordPress user. Anonymous learners cannot have persistent memory.

  • Check that learner identification is working. The generated code tries to identify the learner through these methods (in order):

    1. sm_user URL parameter — automatically added by the [sm_embed] shortcode when the WordPress user is logged in.
    2. SCORM 2004 — reads cmi.learner_id from the SCORM API.
    3. SCORM 1.2 — reads cmi.core.student_id from the legacy SCORM API.
    4. xAPI actor — extracts the learner’s email from the actor URL parameter (e.g., TinCanny).
    5. WordPress session — the HTML Widget calls /whoami which detects the logged-in user via cookie.

    If none of these resolve a WordPress user, memory cannot persist. Check the browser console for identification errors.

  • The SCORM/xAPI identifier is mapped to a WordPress user via the SCORM ID mapping strategy in SlideMind > Settings > Advanced. If your LMS passes an email, username, or custom ID, make sure the mapping strategy matches (default: auto tries all).

  • Check the memory variable’s scope — a variable scoped to module is not visible from a different module. Use course or global scope for cross-module persistence.

How do I view or clear a learner’s stored data?

Section titled “How do I view or clear a learner’s stored data?”

Go to SlideMind > Memory Hub > Learner Data tab. You can search by learner, view all stored variables, and purge data for individual learners or specific scopes.

Does learner memory count against storage limits?

Section titled “Does learner memory count against storage limits?”

Learner memory is stored in your WordPress database (wp_sm_learner_memory table). SlideMind does not impose a limit — the only constraint is your database storage capacity.


RequirementMinimumRecommended
WordPress6.06.4+
PHP8.08.1+
PHP OpenSSLRequired
MySQL / MariaDB5.7+ / 10.2+Latest stable
Outbound HTTPSRequired

SlideMind works locally but not on my hosting provider

Section titled “SlideMind works locally but not on my hosting provider”

Common issues on shared hosting:

  • wp_remote_post blocked or restricted — some hosts disable or limit WordPress HTTP functions. Contact your host and ask if outbound HTTPS POST requests are allowed.
  • Low PHP timeout — AI requests can take 10—30 seconds for complex prompts. If your host sets a short PHP execution timeout (e.g., 30 seconds), long conversations may fail. Ask your host to increase max_execution_time to at least 60 seconds.
  • Firewall blocking outbound connections — some hosts block connections to unfamiliar domains. Ask them to whitelist your AI provider’s domain.
  • PHP OpenSSL not enabled — required for API key encryption. Most hosts have it, but some minimal configurations do not.

Response time depends on several factors:

  1. The AI model — larger models (GPT-4, Opus) are slower than smaller ones (GPT-4o Mini, Haiku, Flash). Consider a smaller model for simple interactions.
  2. Prompt length — longer system prompts and conversation histories consume more tokens and take longer. Keep prompts concise.
  3. Provider load — AI providers sometimes experience high traffic. This is outside SlideMind’s control.
  4. Server location — if your WordPress server is far from the AI provider’s data center, network latency adds up. A server in North America or Europe works well for most providers.

Use the Sandbox’s debug panel to see exact response times and identify bottlenecks.

Yes. SlideMind enforces per-IP rate limiting to prevent abuse:

EndpointDefault limitWindow
/chat60 requests60 seconds
/whoami30 requests60 seconds

The chat rate limit is configurable in SlideMind > Settings > Advanced > Chat rate limit. If a learner hits the limit, they see: “Too many requests. Please wait before trying again.”


Yes. SlideMind uses a proxy architecture — the API key is stored encrypted (AES-256-CBC) in your WordPress database and decrypted only server-side when making a request to the AI provider. The key is never sent to the browser or included in any client-side code.

When a learner sends a message, the following is transmitted to the AI provider:

  • The system prompt you configured for that slide.
  • The learner’s message(s) (and conversation history in Chat mode).
  • Any resolved profile variables referenced in the system prompt (e.g., learner’s first name).

SlideMind does not send: the learner’s IP address, email, WordPress user ID, or any data not explicitly included in the prompt configuration.

All learner data (memory variables, conversation logs) is stored in your WordPress database — not on any external server. You have full control over this data through the WordPress admin.

How do I comply with GDPR / Loi 25 / CCPA?

Section titled “How do I comply with GDPR / Loi 25 / CCPA?”

SlideMind provides several built-in tools:

  • Privacy policy suggestion — auto-inserts relevant text into your WordPress privacy policy page.
  • Data export & erasure — integrates with WordPress Privacy Tools (Tools > Export/Erase Personal Data) so learner data is included in data access and deletion requests.
  • Consent overlay — an optional overlay on the HTML Widget that asks for learner consent before the first interaction. Enable it in SlideMind > Settings > Consent.
  • Memory retention policy — configure automatic deletion of learner memory after 30, 90, 180, or 365 days in SlideMind > Settings.
  • Manual data purge — delete individual learner data from the Memory Hub.

Note: SlideMind is a tool. As the site administrator, you are responsible for your own privacy policy, consent mechanisms, and compliance with applicable laws. See the Security & Privacy guide for full details.


Can I share configurations between WordPress sites?

Section titled “Can I share configurations between WordPress sites?”

Yes. SlideMind supports import/export of courses, modules, and slides as JSON files. Go to the course or slide edit page and use the Export button. On the destination site, use the Import function to load the configuration.

Exported files do not include API keys — you will need to configure the API key on the destination site.

What are templates and how do they differ from slides?

Section titled “What are templates and how do they differ from slides?”

Templates are reusable prompt configurations that act as starting points. When you create a new slide, you can apply a template to pre-fill the system prompt, output variables, and settings. Templates are not tied to a specific course or module.

SlideMind includes 6 built-in templates (3 for Q&A, 3 for Chat). You can also create your own, mark them as favorites, and set a default template.