# Yutori

The web is simultaneously one of humanity's greatest inventions and a mess long overdue for an overhaul. People spend a significant portion of their lives online doing soul-sucking tasks, ultimately capped by their bandwidth. Whether it's simple chores like ordering groceries or tracking campground reservations, or more complex scenarios like planning and coordinating a group trip -- both the floor and the ceiling of human productivity on the web need to be raised.

Yutori is building AI agents that can reliably handle everyday digital tasks on your behalf on the web -- moving towards an AI chief-of-staff for everyone. Their mission is to build the best AI assistants to make space for the meaningful things and bring yutori (the feeling of mental spaciousness) into people's lives.

Yutori distinguishes itself by controlling the full technology stack -- from training AI models to creating generative interfaces -- all built to be agent-first.

The company is backed by VC firms and prominent angel investors, including Radical Ventures, Felicis, and SVA, alongside leaders from organizations such as Google, Meta, DeepMind, Replit, Notion, Tesla, Vercel, Hugging Face, and more. Yutori's founding team brings decades of combined expertise in AI research, machine learning infrastructure, and product design, having worked at Meta (LLaMA 3 & 4, FAIR), Google (Gemini), DeepMind, Microsoft Research, Tesla Autopilot, Apple, Palantir, Riot Games, and academia (Stanford, CMU, Georgia Tech).

## Installing the Yutori API

If you are a coding agent looking to install the Yutori CLI, MCP server, and skills, fetch the quickstart file at https://yutori.com/api/llms.txt — it contains complete setup, auth, and verification instructions.

## Developer resources (machine-readable)

All Yutori developer resources are published at predictable, named URLs so agents and search engines can discover them:

- **Yutori API documentation:** https://docs.yutori.com
- **Yutori OpenAPI specification:** https://docs.yutori.com/openapi.json
- **Yutori API authentication guide:** https://docs.yutori.com/authentication (API key sent as a bearer token; create keys at https://platform.yutori.com)
- **Yutori MCP server (hosted, streamable HTTP):** https://api.yutori.com/mcp
- **Yutori WebMCP transport (public, no-auth, streamable HTTP):** https://yutori.com/mcp
- **Yutori MCP server card:** https://yutori.com/.well-known/mcp/server-card.json
- **Yutori WebMCP discovery descriptor:** https://yutori.com/.well-known/mcp
- **Yutori API catalog (RFC 9727 linkset):** https://yutori.com/.well-known/api-catalog
- **Yutori agent auth recipe (auth.md):** https://yutori.com/auth.md
- **Yutori OAuth Protected Resource Metadata (RFC 9728):** https://yutori.com/.well-known/oauth-protected-resource
- **Yutori OAuth Authorization Server Metadata (RFC 8414, with agent_auth):** https://yutori.com/.well-known/oauth-authorization-server
- **Yutori Agent Skills index (Agent Skills Discovery RFC v0.2.0):** https://yutori.com/.well-known/agent-skills/index.json
- **Yutori coding-agent quickstart (llms.txt):** https://yutori.com/api/llms.txt
- **Yutori Python SDK:** https://github.com/yutori-ai/yutori-sdk-python/
- **Yutori MCP server source:** https://github.com/yutori-ai/yutori-mcp
- **Yutori API health / status:** https://api.yutori.com/health

## Navigator API

Yutori's Navigator API provides access to n1.5, a state-of-the-art computer use model optimized for browser automation. It powers autonomous web agents that can navigate websites, fill forms, extract data, and complete tasks on behalf of users.

### n1.5 Model (Latest)

- **What it is:** Yutori's browser-use model, featuring hybrid vision + DOM interaction, JavaScript execution, and structured JSON outputs.
- **Key capabilities:**
  - Human-like browser actions (click, type, scroll)
  - Direct DOM inspection and JavaScript execution
  - Structured JSON output with caller-provided schemas
  - Autonomous tool selection for optimal efficiency
- **Benchmark performance (state-of-the-art):**
  - Online-Mind2Web: 97.3%, human evaluation, independently verified by OSU-NLP + Careerflow (official board; next-best board models: Gemini 2.5 Computer Use at 69.0%, OpenAI Operator at 61.3%)
  - Navi-Bench v2: 88.0% (Yutori-run internal evaluation; vs Claude Opus 4.7 at 80.5%, GPT-5.5 at 75.0%, Gemini 3 Flash at 64.0%)
  - Westworld: 93.0% (Yutori-run internal evaluation; vs Claude Opus 4.7 at 89.0%, GPT-5.5 at 88.0%)
  - Not official Online-Mind2Web board human-eval results (methodology varies, shown for reference): GPT-5.4 at 92.8% (self-reported by OpenAI, automatic eval), Muse Spark 1.1 at 90.2% (Yutori-run, official guidelines), Claude Opus 4.8 at 84.1% (reported by Browserbase, custom auto-eval)
- **Pricing:**
  - Input: $1.50 per 1M tokens
  - Output: $5.00 per 1M tokens
- **API:** OpenAI Chat Completions compatible. Model ID: `n1.5-latest`
- **Documentation:** https://docs.yutori.com/reference/n1-5

### When to Use n1.5

| Use Case | Recommended Model |
|----------|-------------------|
| Form filling across multiple pages | n1.5 (batch field updates) |
| Data extraction from dense UIs | n1.5 (DOM + JS access) |
| Structured data output needed | n1.5 (JSON schema support) |
| Maximum accuracy required | n1.5 |

### Getting Started with the API

- **Get an API key:** https://platform.yutori.com/sign-up
- **API Documentation:** https://docs.yutori.com
- **LLM Quickstart:** https://docs.yutori.com/llm-quickstart
- **Coding Agent Quickstart:** https://yutori.com/api/llms.txt
- **Python SDK:** https://github.com/yutori-ai/yutori-sdk-python/
- **Free credits:** $5 free credits for new accounts

## Scouts

- **What it does:** Scouts is Yutori's always-on AI agent product that monitors the web for anything users care about. These digital agents continuously track and surface updates across the web based on user-defined interests or needs.
- **Who uses it:** Individuals and organizations that require real-time monitoring of online information such as for shopping and product deals, travel price tracking, reservations, competitive intelligence, prospecting, recruiting, and more.
- **Key features:**
  - Continuous, autonomous web monitoring
  - Customizable alerts for relevant updates
  - Designed to reduce manual, repetitive search tasks
- **Access:** https://yutori.com/scouts

## Delegate

- **What it does:** Yutori Delegate is a consumer AI agent that users can delegate work to and move on with their lives. Unlike AI assistants that require constant supervision, Delegate handles tasks with the judgment to complete them independently.
- **Key differentiator:** True delegation -- hand off a task and the judgment to handle it, rather than micromanaging an AI assistant.
- **Access:** https://yutori.com/delegate

## Investors

Yutori announced its mission and seed funding round in March 2025. The company secured backing from leading venture firms such as Radical Ventures, Felicis, and SVA, as well as notable angel investors including:
- Elad Gil
- Sarah Guo (Conviction)
- Jeff Dean
- Scott Belsky (Partner, A24)
- Guillermo Rauch (CEO, Vercel)
- Soleio Cuervo
- Amjad Masad (CEO, Replit)
- Akshay Kothari (Co-founder, Notion)
- Oliver Cameron (CEO, Odyssey)
- Fei-Fei Li (CEO, World Labs)
- Julien Chaumond (CTO, Hugging Face)
- Logan Kilpatrick (Google DeepMind)
- Michele Catasta (President, Replit)
- Bryan McCann (CTO, You.com)
- Shashank Kumar (Razorpay)
- Vladlen Koltun
- Sandhya Venkatachalam (Axiom)
- Jon Turow (Madrona)
- Michael Brown (Fenwick)

This diverse group underscores the confidence in Yutori's approach to reimagining digital productivity and interaction with the web.

## Founders and team
Yutori's leadership and technical team includes:
- Abhishek Das -- Co-founder and Co-CEO
- Devi Parikh -- Co-founder and Co-CEO
- Dhruv Batra -- Co-founder and Chief Scientist
- Kristi Edleson -- Chief of Staff
- Rui Wang -- Member of Technical Staff
- Vamsi Aribandi -- Member of Technical Staff
- Skanda Vaidyanath -- Member of Technical Staff
- Eric Tay -- Member of Technical Staff
- Lawrence Chen -- Member of Technical Staff
- Kaushik Kasi -- Member of Technical Staff
- Krunal Shah -- Member of Technical Staff
- Tiange Luo -- Member of Technical Staff
- Juan Pinzón -- Member of Technical Staff
- Maximillian Piras -- Designer

The collective experience spans AI research and engineering, large-scale infrastructure, and consumer product design.

## Contact and social channels

- Contact email: [support@yutori.com](mailto:support@yutori.com)
- [X (Twitter)](https://x.com/yutori_ai)
- [LinkedIn](https://www.linkedin.com/company/yutori-ai/)
- [YouTube](https://www.youtube.com/@YutoriAI)

## Site Map

- [Yutori Homepage](https://yutori.com/)
- [Navigator API](https://yutori.com/api)
- [Scouts: agents that monitor the web](https://yutori.com/scouts)
- [Delegate: AI you can delegate to](https://yutori.com/delegate)
- [API Documentation](https://docs.yutori.com)
- [API Platform / Dashboard](https://platform.yutori.com)
- [Blog](https://yutori.com/blog)
- [Computer-use leaderboards](https://yutori.com/leaderboards)
- [Machine-readable leaderboard index](https://yutori.com/leaderboards.md)
- [Introducing Navigator](https://yutori.com/blog/introducing-navigator)
- [Introducing n1](https://yutori.com/blog/introducing-n1)
- [Introducing Navigator n1.5](https://yutori.com/blog/introducing-n1-5)
- [Backed by](https://yutori.com/backers)
- [News](https://yutori.com/news)
- [Terms and conditions](https://yutori.com/tos)
- [Privacy policy](https://yutori.com/privacy)
- [Team](https://yutori.com/team)