Catching up with Weavy v29

What’s new in Weavy

The agents are on a roll, and they’re bringing Weavy even closer to the context of your app.
For the latest updates, we’ve been listening to your requests. Not just around AI, but also around how the Copilot experience shows up across Weavy. You’ve been asking for more control, especially when it comes to customizing the UI, and now we’re delivering. We’re adding more custom element slots, CSS shadow parts, and styling properties to all components. On top of that, we’re spreading the AI functionality throughout, so it’s always close to your app’s context. Just the way you wanted.

v28

This version introduced some bigger changes around the newly released API for AI agents. Everything around agents was renamed from the previous bots naming.

• Renamed the Bots API to Agents API - All names that previously used bot in the name now use agent in the name instead.
• AI agent file search capabilities - Let AI agents work with your files.
• Context data in all components - Now any contextual component can leverage AI data.

v29

This version has some UI adjustments as well as agent improvements.

• Removed the header in Messenger - Let's you provide your own title and buttons properly.
• File annotations for Agents - Let's the agents show what files they are referencing. 
• Configure Agent AI tools - Connect multiple file apps and interface external web services.

"This brings the agents to a new level! It has never been easier to connect your files, realtime data and Web APIS to AI." 

See the changelog for all updates and details on how to upgrade

Renamed the Bots API to Agents API

To keep the APIs consistent and the naming logical, we renamed all the legacy bot naming to agent. This affects mainly the Agents API but also all other APIs where there previously was bots. So if you use any API with anything named bot in it, rename it to agent.

See Agents API in docs

AI agent file search capabilities

The AI agents can now search files for content and knowledge. All this is managed by connecting a file app/component to your agent. The agent will then search the files provided for answers. This means you can pre populate data for your agents. And it can of course be updated as you like, whenever you like. 

Try it out in your Weavy Account.

Context data in all components

The context data property from the Weavy copilot is now available on all contextual Weavy components. This means that you can connect your realtime data to make it available in any of your components. For instance, the AI agents in a Weavy chat component can now know all about the surrounding data you have in your app by just connecting the data property.

See the properties section on each component in UIKit Web

Removed UI header in Messenger 

The header in the top of the conversation list has now finally been removed. Now you can configure a header and title of your own. This makes it easier to align the Messenger with your own UI. We also provided the possibility to customize or remove the create conversation button. You can also inject additional action buttons next to the create conversation button. Additionally, you may also use the new createConversation() method to create new conversations.

See Messenger reference docs

File annotations for Agents

When an AI agent references a file, you now get a list of all the annotated files next to the answers. Preview and work with the files together with the agent, right where you are.

Configure Agent AI tools

You can now configure web tools for the agents to work with. You may request external data or perform custom actions using external API endpoints of your choice. Just add the url and define the purpose and let the agent take care of the rest.

See Configure tools for AI Agents or configure an agent in your Weavy Account.