Paul Omenaca
@houmlandAs the holiday season kicks off—starting with Thanksgiving 🦃 for our US-based users—the Stack AI team is staying in high gear, rolling out new developments and platform enhancements.
Here’s what’s new.
Transform your LLMs into proactive agents by equipping them with tools they can trigger when needed.
Here’s the current lineup:
Select your LLM and add tools using the bottom section. No other inputs are required from your end. Each tool has a settings tab where you can change some parameters (e.g., search location, etc.).
We're continuously expanding this toolset and phasing out outdated plugins for an even better experience.
We’ve optimized workflows to support multiple LLMs being executed at the same time. This is impacting mainly projects with multiple branches of data processing.
Tasks now run simultaneously, cutting latency by up to 10x. For example, a logistics customer saw their latency drop from 135 seconds to just 9 seconds for their project.
You asked, we delivered! Now you can customize the order of inputs and outputs in the form interface.
Here’s how:
It’s that simple—enjoy a more tailored workflow
We’ve listened to your feedback and made functionality blocks more intuitive to navigate in the sidebar:
Data Loaders: Now organized under:
Multimodal Models can be now found on:
We’ve upgraded how knowledge bases work across the platform:
Experience a smarter, more integrated way to manage your knowledge.