sambyte
Forum Replies Created
-
Thank you so much for your kind words! We’re really glad to hear that you’re enjoying the plugin and that it’s making WooCommerce management easier for you.
Regarding the AI Integration, after enabling it, please make sure to click the Save AI Settings button inside the AI Integration section, not the global Save Settings button.
Based on your report, we were able to identify a bug during the initial save. On the first save, the AI settings are incorrectly processed by the global settings handler, which prevents the Generate with AI button from appearing. If you click Save AI Settings a second time, it should work correctly and the button will appear.
Don’t worry, we’ve already identified the issue and will include a fix in our upcoming release.
Thank you again for taking the time to report this and helping us identify the bug. We really appreciate your support!
Hi,
We wanted to let you know that we’ve completed our investigation and have released AskAny v1.15.0, which includes improvements and enhancements to the RAG embedding system.
This release includes updates aimed at improving the reliability and overall behavior of the embedding process based on our investigation. These improvements are intended to provide a smoother embedding experience, including scenarios similar to the one you reported.
Please update to AskAny v1.15.0 and run the embedding process again. After updating, we’d appreciate it if you could let us know how it works on your site.
If you notice any new or unexpected behavior after updating, please feel free to open a new support topic with the details. We’ll be happy to continue investigating and assist you further.
Thank you again for your patience and for helping us improve AskAny!
Thank you for the update.
Based on the screen recording, we don’t see any obvious errors in the network response, and it doesn’t appear that the embedding request itself is failing. At the moment, it’s not entirely clear where the process is stopping.
We’ve also been testing this on our end, but so far we haven’t been able to reproduce the same behavior, so it may be specific to your site or environment. We’ll continue investigating and testing to better understand the cause, and we’ll share any updates as soon as we have them.
One thing you can also check is whether the AI is able to answer questions using your website content. Besides RAG embeddings, AskAny also maintains a separate content indexing system that stores and indexes your site content for AI responses. If the content indexing has completed successfully, the AI should still be able to respond using your website content even if the RAG embedding progress appears to remain at 98%.
Once we’re able to reproduce the behavior or identify the cause, we’ll get back to you with an update.
Thank you for your patience and for helping us investigate this.
Thank you for the additional details. That helps clarify the situation.
Regarding your questions:
- Currently, embeddings are supported only with OpenAI and Gemini. Claude is currently supported for chat responses only.
- At the moment, there is a single API configuration. Having separate providers for embeddings and chat responses is a great suggestion, and we’d be happy to consider adding that in a future update.
As for the embedding process getting stuck at 98%, we’ll investigate this further and perform additional testing on our side to better understand what’s happening.
If possible, could you also share the following?
- Any debug logs, if available.
- The browser’s Network tab output while the embedding process is running. In particular, please look for the request related to the embedding generation, then share the Request Payload and Response. This may help us identify where the process is stopping.
Thank you again for the detailed report. We’ll investigate this further and share an update as soon as we have more information.
Thanks for the report.
Based on our initial findings:
- The selected Claude model (
claude-sonnet-4-20250514) is no longer a current model, which may affect the connection test. We’re updating our supported AI model list to include the latest Claude models. - The embedding process may take longer when processing larger content. We’re working on optimizing the batch processing to improve reliability and performance for larger datasets.
- If possible, please try the same setup using Gemini or ChatGPT/OpenAI and let us know whether you experience the same behavior. This will help us compare the results across different AI providers.
- We’ll continue improving both the supported AI model list and the embedding process in upcoming updates.
Thank you for reporting this and helping us improve the plugin.
Thank you for reporting the issues.
Unfortunately, we’re unable to view the screenshots from the links provided. Could you please either update the sharing permissions so they’re accessible or attach the screenshots directly to this support topic?
To help us investigate further, could you also share:
- Any debug logs or error logs related to the embedding generation or Claude connection (if available).
- The complete error message (if there is one).
- Your WordPress version, PHP version, and the plugin version you’re using.
- The AI provider and embedding model you’re using.
These details will help us better understand the issue and investigate it further.
Thank you!
Thank you, the solution worked.