Octen Debuts Ultra-Fast Search API to Power the AI Agent Era

SAN FRANCISCO, CA – April 22, 2026 – Startup Octen today announced the launch of what it claims is the world's fastest web search API, a move aimed at rebuilding the internet's information retrieval backbone for the burgeoning era of artificial intelligence. Backed by a new $10 million seed funding round, the company is rolling out a proprietary search engine designed not for human eyes, but for the high-frequency, parallel processing needs of AI agents.

With dual headquarters in San Francisco and Singapore, Octen is deploying a global search infrastructure that promises to deliver web data to AI applications with unprecedented speed and scale. The company reports an average response time of around 60 milliseconds and the capacity to handle over one million queries per second (QPS), performance metrics that drastically outpace the architecture of today's human-centric search engines. This new search layer is intended to solve a fundamental bottleneck in AI development, allowing autonomous agents to interact with the live web with the speed and fluidity of accessing their own memory.

Beyond Human-Centric Search

The internet's search infrastructure, dominated for decades by giants like Google, was meticulously engineered for human behavior: a person types a query, scans a list of ten blue links, and clicks one. This sequential, low-volume process is fundamentally incompatible with the way AI operates. An autonomous AI agent tasked with a complex research project might need to issue thousands of queries simultaneously, cross-referencing data points from countless sources in seconds.

This high-concurrency demand breaks traditional search systems. "Search infrastructure was built for humans, ranked by ads and formatted for people to browse. That doesn't work for AI," said Tushar Roy, a Partner at Square Peg, the venture capital firm that led Octen's seed round. "Octen is building the search layer that is faster, cleaner, and purpose-built for how AI consumes information."

Octen's CEO and founder, Kuan Zou, emphasized that speed and concurrency are the foundational pillars for the next generation of AI. "An agent can process thousands of data points simultaneously, which places a massive technical burden on the underlying stack," Zou stated in the announcement. He argues that by achieving sub-100ms latency and million-level QPS, Octen enables a new milestone where AI agents can reason over live, dynamic information rather than being limited by the static, pre-trained knowledge of their models.

Setting a New Performance Benchmark

Octen's performance claims are not just theoretical. The company's proprietary Octen-8B model made waves even before this official launch, sweeping the Retrieval Embedding Benchmark (RTEB) and outperforming models from established industry titans. This benchmark success, which measures precision and long-context understanding, established the company's technical credibility in data retrieval logic just three months after its founding.

The search API itself is built on a proprietary, ultra-large-scale distributed search engine. The reported median response time of 62 milliseconds is more than four times faster than its closest known competitors, a critical advantage for real-time agentic workflows. This speed is complemented by a system designed for high-density workloads, making it the first to publicly claim support for over 1 million QPS at production scale. Underpinning this is a massive, multi-language index covering a trillion-scale portion of the web, structured specifically for machine consumption and updated with minute-level freshness.

This combination of speed and freshness means AI applications can act on breaking news or emerging trends, overcoming the knowledge cutoff limitations that plague many large language models. For example, an AI-powered financial assistant could generate a stock performance report by synthesizing thousands of news articles, regulatory filings, and market data points published just minutes before.

The Crowded Race to Power AI Agents

Octen enters a rapidly evolving and competitive market. A host of companies are vying to become the go-to information source for AI, each with a different approach. Traditional SERP API providers like SerpAPI and Serper.dev offer a layer on top of existing search engines, but this often requires extra steps to clean and process data for AI use, adding latency.

More direct competitors are AI-native platforms like Tavily, which provides summarized context for AI, and Exa, which uses a neural search model and boasts response times under 180 milliseconds. Others, such as Parallel Web Systems and You.com, are also building APIs specifically for AI agent workflows. Octen aims to differentiate itself with its singular focus on raw performance—delivering the cleanest possible data at the highest possible speed and scale.

This fierce competition underscores the critical importance of real-time data for the advancement of AI. As developers build more sophisticated autonomous agents, the quality and speed of their access to the outside world will determine their ultimate capability and usefulness.

A $10 Million Vote of Confidence

The $10 million seed round, with participation from Argor and a cohort of AI scientists alongside lead investor Square Peg, provides Octen with the capital to aggressively scale its vision. The funds are earmarked for expanding its globally distributed server architecture and growing its engineering and developer relations teams.

The confidence of investors stems not only from the technology but also from the team's pedigree. Founder Kuan Zou brings a decade of experience from the front lines of search and AI infrastructure, having previously led Alibaba Cloud AI search and built Baidu's enterprise search platform from the ground up. He has assembled a formidable team of engineers and researchers with experience at Meta, Google, TikTok, Baidu, and DeepSeek.

"What drew us in was the team," noted Square Peg's Tushar Roy. "Zou has put together an impressive team with extensive experience building systems that powered search at significant scale, throughput and reliability. They are among the best in the world at what they do."

Octen is currently in an invitation-only beta, working with design partners across the generative AI ecosystem to power the next generation of autonomous researchers and enterprise assistants. As the agentic internet begins to take shape, the underlying infrastructure that connects AI to the world's information will become one of the most critical and valuable layers of the new technological stack.