DataFirefly SEO AI Blog Pro — Complete guide

Overview

DataFirefly SEO AI Blog Pro turns your PrestaShop 8 or 9 store into an automated editorial publishing machine. The module analyses your catalogue, proposes a complete blog structure in one click, suggests articles in bulk and publishes them on the schedule you define — one article every N days if you wish, with no manual intervention.

Five AI providers are supported and switchable at any time: OpenAI, Anthropic Claude, Google Gemini, Mistral and DeepSeek. Every call is logged with its real USD cost.

Installation

1. Download the module zip from your DataFirefly account.
2. In the PrestaShop back office, open Modules then Upload a module, and import the zip.
3. Click Install. The module creates its 21 tables, its 12 admin tabs, and automatically generates a random security token for the cron.
4. A new “Blog” menu appears in the back-office sidebar.

Upgrading from 1.0.x: simply replace the module folder and reload the page — the upgrade script creates the new tables without touching your existing data.

API key configuration

Open Blog then Configuration. Enter the API key of the AI provider(s) you plan to use:

• OpenAI: key starting with sk- from platform.openai.com
• Anthropic Claude: key from console.anthropic.com
• Google Gemini: key from aistudio.google.com
• Mistral: key from console.mistral.ai
• DeepSeek: key from platform.deepseek.com

Then select the default provider and model. You can switch providers on every generation without reconfiguring.

1-click automatic setup

Open Blog then Strategy & Setup. The “Analyse my shop and propose a structure” button scans your catalogue: product categories, manufacturers, price ranges, a sample of recent products and any existing blog posts.

The AI then proposes:

• 5 to 8 blog categories with name, description, SEO meta and a dedicated expert prompt per category
• 15 to 25 transversal tags
• 1 to 3 author personas with credible bios
• A global editorial line and pillar keywords

Every item comes with a checkbox. Untick what you don’t want, then click “Apply selected items in bulk”. Entities are created immediately. An optional “Editorial focus” field lets you steer the proposal (for example: B2B audience, sustainability angle, technical tone).

Bulk article suggestions

Open Blog then Article suggestions. Choose how many ideas to generate (3 to 50), optionally a target blog category, a funnel mix (TOFU discovery, MOFU comparison, BOFU conversion, or balanced), a target length and a pillar keyword.

Each suggestion is a complete card: SEO title, target keyword, 4-6 H2 outline, suggested category, catalogue products to cross-sell, funnel stage and an AI cost estimate for generation.

Tick the cards you want, then pick a scheduling mode:

• Generate now: articles are processed on the next cron tick or via the manual processing button
• Schedule at a specific date: all selected articles share the same date
• One article every N days: dates are automatically staggered from the start date you set

The “Auto-publish” selector determines whether generated articles are published directly or saved as drafts for review.

Generation queue and cron

Open Blog then Generation queue. Track every job: status (pending, generating, published, failed, cancelled), scheduled date, attempts, real cost and any error message. Cancel, Retry and View post actions are available per row.

For fully unattended operation, copy the cron URL shown at the top of the screen (it contains your security token) and configure it in your hosting cron or a service such as cron-job.org, hourly. On every tick, the cron processes due jobs in batches of 5 by default (configurable), with a maximum of 3 attempts per job on API errors.

The “Process now” button also lets you run a batch manually without waiting for the cron.

Expert mode: per-category prompts

Every blog category has a “Prompt override” field in its edit form. When filled, this text is appended to the AI system prompt on every article generation in that category. You can enforce an expert tone on a B2B category and an accessible tone on a consumer category, within the same store. The automatic setup pre-fills this field with a proposal tailored to each category.

Automatic internal linking

Open Blog then Internal links. Define your keyword-to-URL rules: product, category, CMS page, another post or a custom URL. Per-rule options: rotating anchors, nofollow, exact match, maximum insertions per article and minimum distance between two links. The engine processes each article on save and injects the links into the rendered HTML, with zero front-end performance impact.

SEO, structured data and AEO

The module automatically handles: title and description tags, Open Graph, Twitter Cards, canonical, multilingual hreflang, a dedicated blog XML sitemap, an RSS 2.0 feed and a per-category llms.txt file for AI answer engines. Schema.org BlogPosting, BreadcrumbList and FAQPage structured data is generated on every article. Each article’s FAQ is stored as JSON and rendered as an accessible accordion.

Linked products and product-page widget

Three selection modes for the products featured under each article: manual, automatic by keywords, automatic by AI. On your store’s product pages, an “Articles mentioning this product” widget creates the reverse blog-to-catalogue linking.

Front-end design

The front end follows a refined editorial direction: serif headings, sticky table of contents with scrollspy on desktop, reading progress bar, drop cap on the first paragraph, accordion FAQ, back-to-top button and a print stylesheet. Everything is scoped under the dfblog prefix and compatible with the Classic and Hummingbird themes. All colours and fonts are centralised as CSS variables at the top of front.css for one-line customisation.

Troubleshooting

• 500 error on AI screens: check that the API key of the selected provider is set in Configuration and that your quota isn’t exhausted. The error detail appears in the network response (browser Network tab) and in the Error column of the generation queue.
• Empty or too-short article: the queue marks the job as failed with an explicit message. Check the API key and configured model, then use Retry.
• CSS or JS not loading on the front: make sure you’re on module version 1.1.0 or above, which allows static assets in its htaccess file, then clear the PrestaShop cache (Advanced Parameters then Performance).
• Jobs stuck in pending: check that the cron URL is actually called (opening it in a browser must return JSON) and that the token matches the one shown in the Generation queue screen.

Frequently asked questions

Does the module work multistore and multilingual?

Yes. Posts, categories, tags and authors are multilingual and associated per store. Interface translations ship in French, English, Spanish and German.

Can I switch AI providers along the way?

Yes, on every generation. The five providers share the same internal interface and every call’s cost is logged.

Are generated articles published without review?

Only if you enable Auto-publish. By default, articles land as drafts for review before publication.

What happens on uninstall?

Uninstalling drops the module tables and all blog content. Export or back up your database beforehand if you want to keep the articles.