/w=1920,quality=90,fit=scale-down)
Keywords is building the AI gateway with built-in observability and evals that AI startups use to ship reliable to production.
Respan
Respan AI is a unified AI gateway and observability platform for teams building on hosted LLMs. With just two lines of code, engineering teams can route traffic to providers like OpenAI and Anthropic through a single high-performance gateway, while automatically capturing logs, traces, and metrics in one place. Instead of wiring up separate services for routing, retries, and monitoring, developers get a simple, well-designed control plane that makes it easy to ship and debug AI features in production.
We expect most software teams to keep relying on hosted models for a long time, which means they’ll need a reliable way to sit in front of those APIs. Respan is aiming to become that default gateway by being noticeably better to build with: fast, predictable, and genuinely loved by the developers who use it. We invested because the team has already earned real adoption with a tiny, efficient group, and we believe their focus on developer experience gives them a credible path to winning this layer of the AI stack.
Founders
Andy was previously a Product Design Engineer at Apple, where he worked on AirPods and learned to obsess over every detail of the user experience. He dropped out of the CS program at University of Illinois Urbana-Champaign (UIUC).
Raymond is a systems engineer with a background arge-scale data systems. He studied CS at UIUC with honors and became a ClickHouse expert before co-founding Respan AI. College best friends with Andy.