This startup develops AI agents capable of performing everyday digital tasks on the web, such as ordering groceries or coordinating travel plans. By training their own models and building generative interfaces, they aim to provide users with an AI-powered chief-of-staff for automating online activities.
Funding
$15M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Product
Problem
Individuals spend considerable time online performing repetitive digital tasks, ranging from simple activities like ordering groceries to complex scenarios such as coordinating group travel, which limits overall productivity. Existing AI models often struggle to reliably complete these tasks autonomously due to errors propagating across long sequences of actions.
Solution
Yutori is developing AI agents designed to reliably automate everyday digital tasks on the web, providing users with an AI-powered chief-of-staff. Their agent-first approach focuses on post-training foundation models and implementing a multi-agent system capable of executing multiple tasks and sub-tasks in parallel. By innovating across the entire stack, from training custom models to generative product interfaces, Yutori aims to deliver accurate and dependable AI assistants. The platform employs post-training techniques, including reinforcement learning, test-time search, and model-in-the-loop flywheels, atop open-source models with commercial rights.
Target Audience
Yutori's primary target audience is individuals seeking to automate repetitive online tasks and improve overall productivity with the assistance of AI agents.
Features
- AI agents capable of automating a wide range of digital tasks, from ordering groceries to coordinating travel plans
- Agent-first approach focused on post-training foundation models for enhanced reliability
- Multi-agent system for parallel execution of tasks and sub-tasks
- Custom-trained models and generative product interfaces
- Post-training techniques including reinforcement learning, test-time search, and model-in-the-loop flywheels