Kura AI
Kura AI builds enterprise browser automation and a browser-agent platform that lets teams create production-ready AI agents with natural language rather than traditional scripting. Its stated vision is to turn web interfaces into APIs for the next generation of agentic companies.
At a Glance
- Enterprise software teams
- Developers building agentic companies and web agents
- Organizations needing resilient browser automation
- Businesses in software, productivity, data integration, robotic process automation, security, and enterprise applications
AI Tools by Kura AI
(1)
Kura AI
No-Code AI Browser Agent Builder
Discussions
No discussions yet
Be the first to start a discussion about Kura AI
Latest News
Kura announces 87% WebVoyager benchmark score and claims new state of the art for browser agents, 31% above Claude Computer Use
Kura AI joins Y Combinator Summer 2024 batch as an active San Francisco generative-AI company
Kura documentation update details the Recursion agent, workflow builder, API deployment, observability, and workflow caching
Products & Services
An enterprise browser-agent platform for designing, testing, iterating, and deploying AI agents that interact with websites through natural language. Its architecture uses an initial planner, agent-loop planner, executor, agent-loop critic, and final critic.
A natural-language agent that builds browser-agent workflows: users describe a workflow, approve or modify suggested actions, and watch the agent execute actions while adding corresponding workflow blocks. A beta self-driving mode can automatically confirm and execute actions.
Agents can be deployed as callable APIs or run as applications in Kura's interface; the documentation also describes session-management API endpoints, webhooks, and production deployment.
Market Position
Kura positions itself as a state-of-the-art browser-agent platform differentiated by a multi-agent debate architecture combining planner, executor, critics, vision, and DOM/HTML context. On its published WebVoyager evaluation it reports 87% overall accuracy versus 73.1% for Agent-E and 57.1% for the original WebVoyager baseline; it also reports 90% on a 50-task comparison versus 56% for Anthropic's Computer Use demo. Relevant competitive alternatives include Anthropic Computer Use, Agent-E, WebVoyager-style browser agents, and conventional RPA/browser-automation platforms.
Leadership
Founders
Ronit Basu
Founder and CEO. Y Combinator describes Basu as having four years in ML Ops and prior experience at Google and Convoy.
Darren Hwang
Founder and CTO. Previously spent five years at Meta building security infrastructure and integrity-ranking infrastructure.
Executive Team
Ronit Basu
Founder and CEO
Previously worked in ML Ops at Google and Convoy; Y Combinator says Basu had four years of ML Ops experience before Kura.
Darren Hwang
Founder and CTO
Previously spent five years at Meta on security infrastructure and integrity-ranking infrastructure.
Founding Story
Ronit Basu and Darren Hwang started Kura to build generalized browser agents capable of operating websites through computer vision and DOM/HTML understanding. They argue that important internet functionality remains locked behind human-oriented interfaces without APIs, and that a reliable browser agent can convert those interfaces into API-like capabilities for agentic companies.
Business Model
Revenue Model
Enterprise software and developer-platform model; Kura markets browser-agent building, deployment, API access, support, secure credentials, and enterprise infrastructure through a sales-led enterprise demo. Public materials do not specify whether pricing is usage-, seat-, or subscription-based.
Target Markets
- Enterprise software teams
- Developers building agentic companies and web agents
- Organizations needing resilient browser automation
- Businesses in software, productivity, data integration, robotic process automation, security, and enterprise applications
- Automating repetitive browser-based business workflows
- Web scraping and data extraction
- Automated testing of web applications
- Operating complex web interfaces that lack APIs
- Building proof-of-concept and simple automations with self-driving mode
- Deploying browser agents as production APIs or applications