Skip to main content
Founded in 2025 and headquartered in California and Singapore, Octen is an AI-native search infrastructure company dedicated to building the ultimate retrieval layer for the LLM era. We provide world-class search infrastructure services to developers and enterprises globally, powering the next generation of AI-driven applications and agentic workflows. We offer an end-to-end Web Search API architected specifically for AI agents. By bridging the gap between Large Language Models and the dynamic web, Octen ensures that every agentic query is anchored in the most relevant, real-time, and high-fidelity information available on the internet. Our production-grade performance is industry-leading, featuring a 62ms response time, minute-level index updates, and support for over 1 million QPS. Octen is also redefining SOTA standards in search. Our Octen-Embedding Series has set new global benchmarks, with Octen-8B successfully sweeping the RTEB leaderboard in early 2026. Today, it stands as the industry’s most recognized embedding model, outperforming giants in both precision and long-context understanding. Beyond that, we have built a proprietary, ultra-large-scale distributed search engine, the highest-performing architecture of its kind. As the true infrastructure behind AI, it is engineered from the ground up to guarantee maximum scalability, real-time freshness, and ultra-low latency, meeting the most rigorous demands of AI-driven search scenarios.