Waterfall AI Bulk Product Reviews

Product Review Plugin

Give every product page a review section that feels real: a star-rating summary, photo and video reviews, and post-purchase show-and-tell. The admin panel supports both bulk generation and single-entry fine-tuning, with images and video mixed inline in one and the same review.

No AI API required! Doubao, DeepSeek and Qwen all work for bulk generation

Two ways to add reviews

AI generation (bulk)

Copy the prompt the admin panel gives you → hand it to any AI (ChatGPT, Claude, Doubao — any of them) → paste the result back in and import.

• No AI API to configure: the plugin never calls a model, so there is no API cost at all
• Up to 500 reviews per batch, spread evenly at random across the target products
• Three target scopes: a single product / one category (including sub-categories) / every product on sale storewide
• Optional image and video URL pools, so the AI can hand them out across the reviews as it writes
• Dates are optional, letting reviews carry a sense of time depth; imported reviews land as approved and skip the moderation queue
• Strict format validation: one bad entry rejects the whole batch, so you never get half a batch written

Manual entry (single)

Add one weighty review to a bestseller: text, images and video all inside the same entry.

• One product only: type a keyword and press Enter to search; picking a result locks it as the target
• Rating 1–5, nickname, title, content and date, filled in field by field; the date can be backdated
• Up to 6 images and 1 video, with a preview generated automatically after upload

Storefront

• The review section sits under the product description; both the rating at the top and the Reviews tab jump down to it
• Star summary plus average score, with filtering by star rating and pagination (5 reviews per page)
• Images are auto-compressed and converted to WebP; videos keep their original format. Media is grouped in upload order: images can be enlarged, the video can be played

Admin — reviews in bulk

The list filters by product, rating and status, and supports bulk approve / reject / delete.

Installation

1. After purchase, enable the plugin in the admin panel; it creates its tables automatically
2. Go to Products → Product Reviews and start adding

Limits and rules

AI keywords

Edit the keywords below to match your own industry and requirements, send them to any AI, and let it write reviews in bulk. When you add reviews, paste the output in and its content is used at random for the bulk import.

You are a senior copywriter for a jewellery store. Write 10 real customer reviews in English, following the field format below exactly. Output nothing but the review blocks - no preamble, no explanation.

Product scope: ring

Rules:
1. Each review is 2-4 sentences; 30-80 words in English, 40-120 characters in Chinese, and the content must be at least 10 characters and no more than 2000. Natural spoken tone, like a real buyer. Never mention AI, samples or templates.
2. Mostly 5 stars, some 4, occasionally 3. The text must agree with the rating.
3. The title is 3-7 words and may be left empty.
4. Invent a plausible handle - first name plus initial (e.g. "Emily R.") or a Chinese courtesy name (e.g. 林小姐) - and never repeat one.
5. The date is optional, format YYYY-MM-DD, within the past two years.
6. No order numbers, prices, store names or other products.
7. One block per review, blocks separated by a line containing only ---.

No image or video assets are provided this time; do not output "Image" or "Video" fields.

Output format (keep the field names and their order; the Title, Date, Image and Video lines may be omitted entirely; Content goes last and may span several lines):
Rating: 5
Author: Emily R.
Title: Beautiful setting
Date: 2026-05-12
Content: The ring arrived in a lovely box and the stones really catch the light.
---
Rating: 4
Author: Daniel K.
Title:
Date: 2026-03-02
Content: I wore this ring the whole day and it stayed comfortable,
and the stones kept catching the light.

Output the review blocks only, starting with the first "Rating:" line.