As AI reshapes how we use the internet, a new user persona is emerging: the agent. These AI systems act on behalf of users, from booking travel and paying bills to automating enterprise processes and managing SaaS tools.

Until now, AI models have excelled at producing text and ideas, but when it comes to taking action, they’ve relied on inefficient workarounds: scraping websites, navigating human-oriented UIs, or using manually built plugins. The Model Context Protocol (MCP), which has now been adopted by all major AI players, changes that by offering a secure and structured way for agents to connect to external services. 

MCP unlocks a new way of interacting with the digital world. Instead of perusing a list of flights on a website, you can state where you want to go and an agent finds the best options and books it. Rather than logging into a CRM dashboard, an agent can update a client record and invoice them on your behalf. Human-facing UI is giving way to interfaces for agents.

Today, we’re excited to announce our $6 million pre-seed round to lead this shift, led by Partech, which has yet again thrown their support behind our team and our ideas. We are also thrilled to have the support of K5 Global, Irregular Expression, Yellow, Drysdale, Kima Ventures and Galion.exe, as well as prominent founders in the AI and dev tool space from companies like Mistral, Datadog, and Dataiku. Together they bring expertise in developer platforms, cloud infrastructure, and enterprise software.

Just as the rise of streaming required a new delivery model, the shift to AI agents demands a new foundation. Our team has lived this before when we built video delivery startup Streamroot and helped media companies adapt as millions of viewers moved from broadcast TV to streaming.

The real potential of agents lies in their ability to interact with the digital world around them. For that we need infrastructure built from the ground up for agents, not retrofitted for them. Alpic provides that layer, enabling developers to deploy and manage MCP servers in minutes, with built-in security, analytics, and tooling. The platform greatly reduces operational complexity and streamlines the path to production for agent-accessible services.

This summer, we worked with early customers to deploy dozens of MCP servers. Today, we’re opening our platform in public beta. If you’re building AI-powered services and want to make them agent-accessible, try it out now: https://app.alpic.ai!

We couldn't be more excited for this new chapter in defining the infrastructure for agent-first computing. Thank you to our investors, partners, and all those who believed that we could make this possible.