Dedalus Labs

Dedalus Labs is a cloud platform for developers creating agentic AI applications. Founded in 2025 by Catherine Di and Windsor Nguyen, the San Francisco-based company is part of Y Combinator's Summer 2025 batch. The founders identified a market gap while building a stateful code execution sandbox, noting the difficulty of connecting large language models (LLMs) with tools even before the advent of the Model Context Protocol (MCP).

The company provides a drop-in MCP gateway that connects any LLM to any MCP server, whether local or fully managed by Dedalus Labs. The platform is designed to simplify the development process, reducing setup time from weeks to minutes by handling configurations, hosting, scaling, and smart model hand-offs. This allows developers to avoid using Docker or YAML files. The core service offers an OpenAI-compatible SDK that lets users orchestrate complex agentic workflows through a single API, enabling them to route between any LLM provider and connect various tools from a hosted marketplace.

Key features include Universal Model Access, which allows users to switch between models like GPT-4, Claude, and Gemini with a single line of code, preventing vendor lock-in. The platform also supports Hybrid Tooling, combining users' Python functions with pre-built MCP tools from a marketplace for capabilities ranging from web searches to data analysis. It enables serverless scaling and instant production with hot-reloading for logic and tools, ensuring zero downtime. A future feature will be a marketplace for developers to monetize their agents or servers, with plans for an 80% creator share and instant payouts.

Keywords: MCP gateway, agentic AI, AI workflows, large language models, developer tools, cloud platform, API endpoint, serverless, workflow automation, AI agents, model context protocol, AI development, multi-model AI, hybrid tooling, SDK, devops