AI Chatbot for Magento 2 Documentation

Overview

AI Chatbot for Magento 2 is a customer-facing product assistant for merchants who want 24/7 conversational product discovery. It adds a responsive floating widget, enriches prompts with storefront catalog context, returns relevant product cards, and gives administrators a searchable conversation log.

Requirements & compatibility ★

Confirm your platform and deployment workflow before purchasing.

Installation

Composer (recommended)

1. Get your private repository access.Sign in to your NSPlug customer account and copy the Composer authentication instructions shown for your purchase. Keep the token in Composer authentication storage outside the Magento repository; never commit credentials or auth.json.
2. Add the NSPlug repository and review the install.Run these commands from the Magento project root. The released package is nsplug/aichatbot and installs to vendor/nsplug/aichatbot.
   composer config repositories.nsplug composer https://www.nsplug.com/composer
   composer require nsplug/aichatbot:^1.2 --no-update
   composer update nsplug/aichatbot --with-dependencies --dry-run
   composer update nsplug/aichatbot --with-dependenciesCopy
3. Deploy the Magento module.Enable NS_AiChatbot, then run setup upgrade, dependency compilation, static-content deployment, and cache flush using the commands below.

Manual / ZIP (alternative)

1. Extract to the exact module path.Place the contents at <magento-root>/app/code/NS/AiChatbot. The directory must contain registration.php, not another nested package folder.
2. Register the manual namespace and dependencies.Add "NS\\AiChatbot\\": "app/code/NS/AiChatbot/" to the Magento root autoload.psr-4 mapping, confirm the root project satisfies the runtime requirements in the module’s composer.json, and run composer dump-autoload -o.
3. Run the Magento setup commands.Run them as the filesystem owner—not root—using the PHP binary for the store.
   bin/magento module:enable NS_AiChatbot
   bin/magento setup:upgrade
   bin/magento setup:di:compile
   bin/magento setup:static-content:deploy -f
   bin/magento cache:flushCopy
4. Configure the provider.Go to Stores → Configuration → General → AI Chatbot → Settings.

Upgrading to version 1.2.1

Back up the module, database, and configuration, then replace the manual-install files with the 1.2.1 package. When upgrading older module code, run the setup, compilation, and static-content commands above as part of your normal deployment process. The upgrade patch clears legacy shared answer-cache entries. Purge old FPC/Varnish/CDN pages so shoppers receive the updated widget.

After the final cache flush, save AI Chatbot configuration at the applicable scope or allow the next scheduled license refresh, then verify the widget and Test AI Prompt. Magento cron must be running. The first load clears older browser chat transcripts once. Every eligible question and retry now makes a fresh provider request, which can increase usage compared with earlier cached replies.

Verify installation

Expect “Module is enabled” and value 1. Load a storefront page, find the floating chat toggle, and use Test AI Prompt after saving credentials. If the module is absent, confirm the folder case, filesystem permissions, Composer autoload, and registration.php; then rerun composer dump-autoload and setup:upgrade.

License activation

1. Copy all three values from your purchase email/account.You need the license key, license token, and license server URL.
2. Open the extension settings.At the correct website scope, enter the values under Stores → Configuration → General → AI Chatbot.
3. Save and verify the license.Saving AI Chatbot configuration refreshes a signed license for the selected storefront domain. Storefront requests use the verified license cache; the background job refreshes it every five minutes. A full cache flush clears cached licenses, so save configuration again or allow the next scheduled refresh before checking the widget. To move domains, deactivate the old production domain in your customer account, then save configuration on the new one.

Download the exact source archive for any version licensed during your term from Account → Downloads → AI Chatbot. Keep the archive with your deployment artifacts.

AI provider setup

OpenAI setup ★

1. Create a project.Sign in to the OpenAI Platform, open the project selector, and create or select a dedicated production project.
2. Add billing and a budget.In organization billing, add a payment method. In project limits, set a monthly budget and notification thresholds.
3. Create a secret key.Open API keys → Create new secret key, restrict it to the project, copy it once, and store it in a secrets manager.
4. Save and test.Select OpenAI in Magento, enter your provider-issued API key, select a model, save configuration, then click Test AI Prompt.

Anthropic setup ★

1. Create or select a workspace.Sign in to the Anthropic Console and use a dedicated production workspace for the Magento storefront.
2. Configure billing and limits.Add billing credits or a payment method, then review the workspace spend limit and the organization rate limits before launch.
3. Create an API key.Open Settings → API keys, create a key for the production workspace, copy it once, and store it securely. Use a standard workspace API key—not an Admin API key.
4. Save and test.Select Anthropic in Magento, enter your provider-issued API key, select an available Claude model, save configuration, then click Test AI Prompt.

Google Gemini setup ★

1. Create or import a project.Sign in to Google AI Studio, open Dashboard → Projects, and create or import the Google Cloud project dedicated to the Magento storefront.
2. Create an API key.Open Dashboard → API Keys, create a key for that project, copy it once, and store it securely. Restrict the key to the Gemini API when the option is available.
3. Configure billing and usage controls.For production or higher limits, use Set up billing to link Cloud Billing and fund the paid tier. Review usage in the AI Studio dashboard and configure billing alerts for the linked project.
4. Save and test.Select Google Gemini in Magento, enter your provider-issued API key, select an available Gemini model, save configuration, then click Test AI Prompt.

On timeout, API failure, rate limit, or exhausted quota, the widget displays an error or fallback response and may still show locally matched products. Failures are recorded through Magento logging. The next question makes a fresh provider request. There is no automatic failover to another provider; configure its key and select it in Magento when switching.

Configuration reference ★

Catalog knowledge & storefront operation

Refresh and corrections

1. Correct the Magento source data.Update product text, attributes, stock, category, CMS content, or sitemap—not the generated answer.
2. Reindex and clean cache.bin/magento indexer:reindex && bin/magento cache:clean. Each question reloads eligible catalog matches and current-product details through the answer endpoint. The widget sends one answer request per question; the separate product-search endpoint remains available for custom integrations. Answers and temporary provider failures are not cached or shared, so every eligible question and retry makes a fresh AI request. Sitemap URL context remains cached for 12 hours. If you perform a full cache flush, refresh the chatbot configuration or allow the next license cron run.
3. Constrain behavior.Add verified policy and escalation instructions to the System Prompt. Review Marketing → Chatbot Logs, delete obsolete rows, and retest in an incognito session.

For human escalation, add a support URL, phone number, or email to the System Prompt, for example: “If confidence is low or the shopper asks about an order, link to /contact and do not guess.” There is no native agent queue.

Placement, styling, language

The widget is injected through view/frontend/layout/default.xml; product context is added on catalog product pages. Override templates in app/design/frontend/<Vendor>/<theme>/NS_AiChatbot/templates/, layout in the theme’s NS_AiChatbot/layout/, and CSS/JS via standard Magento theme fallback. Use the admin color pickers for no-code changes: the header color also controls the send button, while the header font color controls both header text and the close icon. The header title, header avatar, and floating toggle icon can each be customized independently; empty image fields use the bundled defaults. Set tone/language in System Prompt; configure each store view independently for localized greetings and policy.

Security & privacy ★

Performance

Each provider request occupies one PHP worker while awaiting the provider and performs catalog/context preparation plus one log insert. On shared or low-memory hosting, start with a fast model, 8–12 product links, 5–8 categories, and a short prompt; cap concurrent traffic upstream and ensure at least 512 MB PHP memory according to your Magento release requirements.

Ten common problems ★

For developers

Extend with a theme override, layout XML, DI preference/plugin, or observer in a separate module. Prefer a plugin around public methods and a dedicated frontend integration over copying controller logic. Do not edit files under vendor/nsplug/aichatbot; Composer upgrades replace them. Keep overrides narrow and regression-test prompt data, escaping, CSP, CSRF, caching, and provider errors.

Upgrading & uninstalling

Upgrade

Database configuration survives a normal upgrade. Back up code, database, composer.lock, and config before starting. Review breaking changes and test provider calls, the widget, CSP, and store scopes in staging.

Downgrade

There is no automated downgrade path. Restore the pre-upgrade database and composer.lock together; do not install old code over a migrated database unless release notes explicitly permit it.

Disable or fully remove

Remove app/code/NS/AiChatbot instead of composer remove for a manual install. Magento config values remain unless explicitly deleted from core_config_data where path LIKE 'aichatbot/%'. The table/config deletion is permanent; export logs and back up first.

Changelog ★

Product details: View product page ↗. Release dates describe this package’s documentation baseline; validate available downloads in your account.

Support

Included: installation defects, reproducible extension bugs, documented compatibility, license/download access, and clarification of shipped features. Paid customization: theme/headless/Hyvä builds, prompt or catalog strategy, data migration, provider account management, integrations, custom endpoints, performance tuning, and changes to third-party code.

Review the refund policy ↗ before purchase. Compatibility or scope questions are best resolved before ordering—send your exact Magento, PHP, theme, and hosting versions.