Prestashop Ecommerce AI Assistant Test

51 views
Skip to first unread message

diego garcia

unread,
Jun 17, 2026, 11:01:28 AMJun 17
to Chrome Built-in AI Early Preview Program Discussions
Hi  Developers!
Module test in Prestashop (old version) in a version for stagging and testing.
The module has been programmed with Vibe Coding to see the capabilities of Nano with an Ecommerce (1333 products in test).
Due to the limitations of Nano in the browser, I'm still reviewing the best way to pass the data to it so that it has proper context, because it boggles a little or reaches its limit.
1) Install the Module
2) In the Prestashop BackEnd, check that Chrome's built-in AI capabilities are active
3) Temperature and Top_K can be set
4) Important: I generate a file that is the knowledge base for the model in json, (1333 products) I am testing so that it also includes the CMS pages, and files in .txt or .md format
5) I am reviewing limitations for more than 5000 products.
6) On the front-end, a floating wizard starts, with a color indicator (whether AI is available or not)

It would be good to know the capabilities of the model, and what is the best way to give it context with the least use of tokens so as not to overload it and have an accurate response.
I think a numerical value for temperature and top_K is better than properties, since these values are the standard currently.
Given that much navigation is through a smartphone, will it be possible for integrated AI to have similar performance?
Regards
Ecommerce-Asistant_Example.gif
PrestashopIA_Module_Local_Gemini_Nano_4.png
PrestashopIA_Module_Local_Gemini_Nano_5_Search_Index.png
PrestashopIA_Module_Local_Gemini_Nano_2.png
PrestashopIA_Module_Local_Gemini_Nano_3.png
PrestashopIA_Module_Local_Gemini_Nano_1.png

Thomas Steiner

unread,
Jun 17, 2026, 11:10:38 AMJun 17
to diego garcia, Chrome Built-in AI Early Preview Program Discussions
Hola Diego,

Regarding your question on session management, I've documented some best practices in an article. Regarding the parameters, we're actually moving to sampling parameters for better model interoperability:

enum  { , , , ,  };

On mobile, we currently don't support the Prompt API, but we've definitely heard demand for it. The recommendation is to use Firebase AI Logic with "prefer_on_device".

Cheers,
Tom 

--
You received this message because you are subscribed to the Google Groups "Chrome Built-in AI Early Preview Program Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chrome-ai-dev-previe...@chromium.org.
To view this discussion visit https://groups.google.com/a/chromium.org/d/msgid/chrome-ai-dev-preview-discuss/7418c7f1-c7a7-4ea5-882e-e8feee01867cn%40chromium.org.


--
Thomas Steiner, PhD—Developer Relations Engineer (blog.tomayac.comtoot.cafe/@tomayac)

Google Spain, S.L.U.
Torre Picasso, Pl. Pablo Ruiz Picasso, 1, Tetuán, 28020 Madrid, Spain

CIF: B63272603
Inscrita en el Registro Mercantil de Madrid, sección 8, Hoja M­-435397 Tomo 24227 Folio 25

----- BEGIN PGP SIGNATURE -----
Version: GnuPG v2.4.8 (GNU/Linux)

iFy0uwAntT0bE3xtRa5AfeCheCkthAtTh3reSabiGbl0ck
0fjumBl3DCharaCTersAttH3b0ttom.xKcd.cOm/1181.
----- END PGP SIGNATURE -----

Kenji Baheux

unread,
Jun 17, 2026, 10:50:11 PMJun 17
to diego garcia, Thomas Steiner, Chrome Built-in AI Early Preview Program Discussions
Thanks for sharing Diego!

Is this something you have written about on your blog, or might consider writing about?
Curious to learn more as you keep exploring this.



--
Kenji BAHEUX (my how-to)
Product Manager - Chrome
Google Japan
Reply all
Reply to author
Forward
0 new messages