After much back and forth, we decided to stop the closed beta on metadata for dynamic pages because:
it wasn’t being used that much by the beta testers.
with competing priorities, we couldn’t find space on our 2024 roadmap for new SEO features.
That said, we realize there are use cases where SEO is paramount (e.g. marketplaces) and have put together a tutorial that shows you how to use Cloudflare as a reverse proxy to have dynamic metadata on the dynamic pages of your WeWeb apps.
It’s a little bit more on the technical side so if you struggle with this, don’t hesitate to create a new topic in the community, sharing screenshots or a video of what you tried, and where you’re stuck.
And if you’re wondering which features WeWeb is prioritizing in 2024, the replay of the live stream in which @Raphael answers a bunch of user questions and presents the roadmap priorities.
Kinda sad to see AI prioritized over a proper SEO solution, especially after having personally fought for it quite a lot, but completely understandable.
Quick question, could this tutorial be applied to contents other than metadata too? could we render all the contents on page by SSR following similar step on this tutorial?
Yep, I completely understand. At this stage, we feel that AI-features will help more users and it’s where I energy can have the most impact with the resources available in the short term.
We had dozens of people in the closed beta for metadata of dynamic pages and only 2 are using the feature in production. On the other hand, WeWeb Copilot is used multiple times a day, everyday by a large portion of our users.
Yes, absolutely. We think we can take this one step further to implement SSR on all the content in a page We haven’t had a chance to explore the solution fully though so can’t guarantee it 100%.
Were all the websites in the Beta leveraging dynamic content? Because I can’t see how or why you wouldn’t want to use it if you have dynamic content. I for one, would love to have the dynamic metadata as the product pages can’t have dynamic titles out of the box which is a shame.
The AI works sometimes, but it’s honestly not that reliable in my experience. I’m more productive using chatGPT or Anthropic direct, copying in the code and having a full conversation, explaining I’m using weweb and pasting in the URL to the Docs. Rarely is the first response correct, and therefore, the ongoing conversation to improve and iterate is important. The ability to have ongoing conversation is missing at present.
We haven’t had a chance to explore the solution fully though so can’t guarantee it 100%.