Аккаунт в General iT
Войдите или создайте аккаунт для управления услугами в личном кабинете.
Страны не найдены. Попробуйте другое название или код.
MODULE FOR 1C-BITRIX
Denis Spiridonov
The module accelerates website content population by enabling mass generation of data for product cards and sections using neural networks. Since the solution operates on information blocks, its capabilities are not limited to catalogs; you can generate any content for elements and sections of any information block.
The module accelerates website content population by enabling mass generation of data for product cards and sections using neural networks. Since the solution operates on information blocks, its capabilities are not limited to catalogs; you can generate any content for elements and sections of any information block.
Key functions include filling properties and fields of information block elements and sections, such as descriptions, SEO fields, standard properties, and user-defined fields (UF) for sections. Users can select a preferred provider or use an OpenAI-compatible service. Supported providers include GigaChat, Yandex AI Studio, and DeepSeek. Universal platforms with a wide selection of foreign models are also available, including Polza.ai and Openrouter. The module does not charge fees for neural network operations; costs depend on the tariffs of the selected provider.
An AI assistant is available in the module settings to help define generation rules without manual input. You can specify content requirements via a dialog window, after which the request is sent directly to your chosen provider, ensuring process transparency.
The module provides product image generation. It can recognize and describe photographs, as well as generate new product images, provided the selected provider supports the relevant models.
Profiles. The system allows creating an unlimited number of profiles. Each profile is linked to a specific provider and a filter for products or sections. This enables generating product content in one profile and blog content in another, with all settings remaining completely isolated.
The result is a flexible tool for creating both text and graphic content. To understand the module configuration process, refer to the example of a step-by-step quick setup.
General iT can handle the implementation of “AI Content Generation: filling fields, properties, and SEO for information blocks using AI”, from installation to configuration and compatibility checks.
UPDATE HISTORY
The main feature in this version is image generation.
Security:
Protection has been strengthened when downloading images via a link.
New:
Pipeline step "Image": a neural network creates or redraws product images (for example, by changing the background) and saves them to the preview image, detailed image, or the "file" property.
Image generation via Polza.ai, OpenRouter, Yandex AI Studio, GigaChat, and OpenAI-compatible services (including gpt-image and DALL-E).
Product photos can be passed to a text step—for example, to generate a description based on a photo.
Example images in the step allow the neural network to maintain the required style.
In the element card, generated images can be compared as "before/after," accepted, rejected, or rolled back; previous versions are preserved and any can be restored.
The AI assistant helps configure the image step, and such a step can be tested on the "Testing" page before launching.
Improved:
New layout for the element list and element card: search, filter by sections, statuses "In Progress" and "Awaiting Generation," and generation progress by step.
Configurable number of generations per run and timeout for long-running generations.
Pre-check of the pipeline warns about unsuitable models, formats, or image counts; such elements are skipped without consuming tokens.
Settings have been regrouped on the "Main" page, and the step editor has become more compact.
Fixed:
Fixed errors in processing provider responses and displaying element statuses.
New:
In the LLM provider settings for Yandex AI Studio, a step-by-step guide with screenshots has been added: how to create an API key and where to find the catalog ID.
In the LLM provider settings for GigaChat, a step-by-step guide with screenshots has been added: how to create a GigaChat API project, obtain an authorization key, and select the API version.
Improved:
In the model selection table, the columns for input and output prices have been narrowed: the model code and name no longer shrink or wrap.
Fixed:
Fixed the names of the module menu items in the admin panel.
New:
The module interface has been completely redesigned: a new generation dashboard, an "Elements" page with filters and confirmation of proposed values, a new pipeline and step page, an updated prompt collection, and a dark theme.
Step AI assistant: describe the task in words, and it will suggest a prompt, a write field, and a validation field.
Ready-made step presets ("Product Description", "SEO: title and description", "Translation", "Gathering facts with web search", etc.) and the ability to save your own presets.
Prompt generation based on a description and a full-screen prompt editor with field insertion via "@".
Intermediate steps without writing to a field: their results are passed to subsequent steps.
New providers: DeepSeek (with web search), GigaChat, and Yandex AI Studio.
Model reasoning control: for the entire module and individually for each step.
OpenAI-compatible provider: a list of models from the server, selection of the authorization method, custom HTTP headers, timeout, and SSL verification — for Ollama, LM Studio, vLLM, Azure, and others.
Improved:
A new, more reliable generation queue: elements are not duplicated, stuck tasks are returned to work, and on error only the failed step is retried — without unnecessary costs for repeats.
On the dashboard: neural network usage for 7/30/90 days, auto-refresh, the reason for agent stoppage, and an "Enable" button.
In the element card: "Accept", "Reject", or "Revert" for each field.
Proxy is now shared for all providers.
Fixed:
Empty and SEO-condition profile filters no longer return the entire information block for generation.
Generation revert correctly restores multiple and HTML properties.
Retry after failure no longer writes the result twice, and tasks for deleted elements are removed from the queue.
Many minor interface fixes.
Security:
HTML from the neural network is cleaned of scripts before writing to elements and sections.
Regular expressions in text checks are validated before saving and execution.
Errors in settings no longer expose technical details.
- Module renaming
- Security improvements
- Fixes in demo version operation
- Increased the demo generation limit of the module to 500 generations
- Added the ability to generate content for information block sections
- Added additional response time wait settings
- Improved pipeline validation before launch; the module can automatically terminate the pipeline if continuation makes no sense (negative provider balance or invalid pipeline settings)
- Added the ability to use multiple properties in prompts (inserted via commas)
- Minor improvements
USER EXPERIENCE
How does this solution work in a real project? Share your experience and help others choose.
There are no reviews of this solution on git.ru yet. Tell us what worked well and what could be improved.
Verified purchaseAdd a key if you wish. A verified review gets a badge, priority placement and more rating weight.
Open to every clientJust sign in to your account. A key is optional; all reviews are moderated.
YOUR PROJECT INFRASTRUCTURE
FROM THE SAME DEVELOPER
Denis Spiridonov
The module provides reliable synchronization of product data from MySklad to the 1C-Bitrix website management system. It is a ready-made solution that allows you to configure the data exchange once and then operate in automatic mode without constant user intervention. The tool is optimized for catalogs containing a large number of products and their modifications.
The module provides flexible two-way order integration between the 1C-Bitrix CMS and the MoySklad service. The "Online Store + CRM" edition is supported. Important: The module operates only on paid tariffs because the free tariff lacks webhook functionality. Key features: Export orders to MoySklad. Import orders to the 1C-Bitrix website.
The module fills the gap in order automation for 1C-Bitrix Small Business and Business editions, which lack the built-in Robots functionality available only in versions with CRM. This enables flexible configuration of order behavior according to predefined scenarios, a capability absent in standard tools.
MORE FOR YOUR WEBSITE
The module automates image compression on the website in accordance with Google PageSpeed Insights recommendations. It supports integration via CDN OptiPic, allowing optimization without modifying image URLs. The main advantage of CDN integration is the automatic conversion of images to the modern WebP format to minimize file size. Browsers that support WebP receive images in that format.
The module is designed for online stores running on the 1C-Bitrix platform to automate the integration of supplier product catalogs. It is also suitable for news and city portals to facilitate content population. The solution supports data import in Excel, XML, YML, CSV, and RSS formats.
The module is designed for the «1C-Bitrix: Site Management» platform and enables importing infoblock elements, including products, from Excel files. It also supports data loading into Highload blocks.
BEFORE YOU ORDER
