Changelog
Stay up-to-date with the latest updates to Yutori's products.
The multi-agent architecture powering Scouts
An average Scout run invokes 76 agents across a wide variety of tools, and processes ~1M tokens, for just 35 cents per run — all in service of delivering timely, comprehensive coverage of information across the long tail of the web.
We published a deep dive into the engineering principles behind Scouts — the architectural decisions that make continuous monitoring practical, the pitfalls we encountered, and how we balanced quality with cost.
Scouts is now generally available
Today, we're making Scouts available to everyone.
With a simple query, Scouts lets you deploy a team of AI agents to monitor anything for you — for house hunting, early stages of travel planning, sourcing leads, discovering rare products, job search, staying on top of niche news / research / podcasts, discovering local events, etc.
The underlying agent architecture is incredibly powerful — subagents all the way down, powered by our own web navigation agent, and with access to way more tools than before.
This what the future of interfacing with the web looks like. Where you're not sitting there manually browsing and refreshing, buried in tabs, ads, noise, distractions, context switches.
Scouts run 24x7 in the background on the web. So you have the mental space to focus on what's most meaningful to you.
Private Scouts
Scouts can now be made private so only you can view them.
Public Scouts are accessible via public-but-unlisted URLs, and can be shared with others for them to subscribe to. In either case, only the creator can edit the Scout.
Introducing Navigator and our API
Today, we're releasing Navigator — a state-of-the-art web agent that autonomously navigates websites to complete everyday tasks — available via our API.
It achieves 78.7% on Online-Mind2Web and 83.4% on Navi-Bench, with per-step latency 2-3x faster than other computer-use models like Claude 4.5, Gemini 2.5, and Claude 4.0.
Navigator can handle a wide range of real-world web tasks — checking availability, comparing prices, filling out forms, making reservations, ordering food, completing purchases, etc.
The bitter lesson for web agents
The last 1 year of building web agents has taught us a new bitter lesson that we think not many others are internalizing yet.
Agents that perceive the web like humans — using screenshots of websites — navigate and generalize better than agents that rely on the HTML or DOM.
Support for replying to Scout emails to give feedback
You can now reply directly to any report email to provide feedback to your Scout. This feedback gets incorporated into the Scout's subsequent runs, making it better at finding exactly what you're looking for.