Over the past few months, some WeWeb AI requests on large applications returned errors like “Selected content too large” or “Input is too long for requested model.”
With this update, those issues should occur far less often thanks to two key improvements:
• Smarter context selection
• New AI modes
Check out what's new 👇
We’ve revamped the context selection menu in the WeWeb AI chat to give you more control over what’s included in each call.
Instead of automatically pulling in everything (workflows, formulas, popups, etc.), you can now choose exactly which parts you want to include as context:
You’ll avoid unnecessary noise and keep AI outputs more accurate!
Check out our docs to learn more.
WeWeb AI now includes six distinct modes, giving you more control over interactions with your app 👇
• Build mode: edit pages’ visual elements and logic. Manage everything that is on the page, including workflows, formulas, and overall page structure.
• Ask mode: chat with WeWeb’s documentation and get guidance on features, best practices, and specific tasks.
• Coded component mode: edit existing coded components or generate new ones from scratch.
• Supabase mode: manage your Supabase integration, including creating tables, setting up RSL policies, and configuring edge functions.
• Formula mode: generate and edit formulas (triggered automatically when editing formulas).
• Workflow mode: generate and edit workflows (triggered automatically when editing workflows).
Conversations started in one mode can carry into other modes, with your context preserved:
Check out our docs to learn more.
We’ve improved the process of creating AI-generated workflows. Instead of always generating new ones, the AI will now reuse existing workflows when possible, saving you time and reducing rework.
Your existing workflows remain intact, so you won’t lose the work you’ve already done.
AI-generated workflows are now no-code by default, with minimal JS where necessary.
That's all for now! As always, let us know what you think 🙂
