Feedoptimise moves AI feed optimization onto its own GPUs and adds a free monthly AI quota to every plan

MILTON KEYNES, UNITED KINGDOM, October 28, 2026 /EINPresswire.com/ — Feedoptimise, a product feed management and optimization platform, has moved its free AI models from a cloud provider to GPUs in a data center it manages. With the move, Feedoptimise runs Alibaba’s open-weight Qwen 3.8 model, gives users full control over the prompt, adds image prompts and includes free AI requests in every plan, equal to twice the plan’s monthly product limit. Feedoptimise believes it is the first product feed management platform to run AI models on GPU hardware it operates itself.

Feedoptimise added its first self-hosted open model, Qwen 2.5, in January 2025. Since then, the team ran the models it could rent GPU capacity for, at prices the cloud provider set. With its own hardware, Feedoptimise has predictable capacity and costs, and the team can upgrade to newer open models on its own schedule. The same GPUs run BiRefNet, an open-source image segmentation model behind the Feedoptimise Image Editor, and Feedoptimise plans to add more open-weight models to its self-hosted lineup.

Retailers and agencies get four changes:

1) A newer model. Qwen 3.8 replaces Qwen 2.5. Newer Qwen generations follow detailed instructions better and handle more languages, so generated titles keep to the retailer’s rules and extracted attributes come back in a predictable format.

2) Full prompt control. With the new Feedoptimise-AI Prompt modifier, users write the whole prompt and reference any attribute in their feed. Typical uses include rewriting titles to a set pattern, writing channel-specific descriptions, mapping product types to a channel taxonomy, filling missing attributes such as material or gender, scoring title quality and translating content for new markets.

3) Image prompts. Qwen 3.8 reads images as well as text. The Feedoptimise-AI Image Prompt modifier sends one or more product images to the model with the prompt and returns a single value, a sentence or structured JSON. Retailers can use it to tell lifestyle shots from model shots, or to extract color and material from product photos.

4) A free monthly quota. Every plan includes free AI requests equal to twice its product limit, so a 10,000-product plan gets 20,000 requests a month. Feedoptimise caches results and reprocesses a product when its source data or settings change, so unchanged products use no requests. Users who need more can raise their monthly limit for a small additional fee, and pay it in the months when they use the extra requests.

Retailers can use the free model to prepare product content for each channel they send feeds to through Feedoptimise, including Google Shopping, Microsoft Advertising (Bing Shopping), Meta and Amazon. That includes AI shopping assistants such as ChatGPT and Microsoft Copilot, which recommend products from a product feed instead of reading the retailer’s website, and ChatGPT ads, which run from the same feed.

Feedoptimise runs the self-hosted models on a privacy-first basis. The models work with the product data retailers already send to Feedoptimise in their feeds and nothing else, and Feedoptimise processes that data on servers it operates. Feedoptimise does not share it with any external AI provider or use it to train AI models. Users also need no API key or separate AI billing account.

Users who prefer commercial models can connect ChatGPT, Gemini, Claude, Grok or DeepSeek with their own API key, or more than 400 models through OpenRouter, and compare results against the free model with Feedoptimise’s built-in A/B testing. Existing users of the AI Title Optimiser, AI Descriptions Creator and AI Attributes Extractor keep their current setup, and all three modifiers run on Qwen 3.8 starting today.

Availability
All Feedoptimise customers can use the self-hosted Qwen 3.8 model, the Feedoptimise-AI Prompt modifier and the Feedoptimise-AI Image Prompt modifier starting today, within their free monthly quota. Retailers can try them on a free trial.

About Feedoptimise
Feedoptimise is a product feed management and optimization platform for online retailers, brands and agencies. It imports product data from Shopify, Magento, WooCommerce, Salesforce and other sources, and lets users edit, enrich and test that data with rules, AI models and A/B tests before sending it to Google Shopping, Microsoft Advertising (Bing Shopping), Meta, Amazon, ChatGPT, Microsoft Copilot and other channels. Feedoptimise also offers an AI Feed Agent and an MCP server that connects product feeds to Claude and ChatGPT. The company provides self-managed and managed service plans and has operated since 2011. Learn more at www.feedoptimise.com

Marcin Rosinski
Feedoptimise

Media gallery