A San Francisco- and Zurich-based startup, Manufact, has secured $6.3 million in seed funding to build infrastructure for a future where artificial intelligence agents, not humans, are the primary users of software. The company’s core idea: as AI takes over routine digital tasks, every application will need an interface tailored for machine-to-machine interaction. This is a fundamental shift in how software is designed and used, and Manufact aims to be the key enabler.
The Rise of Agentic Software
For decades, software has been designed with the human user in mind. Buttons, menus, and dashboards translate human intent into machine action. But the landscape is changing. AI agents are increasingly handling tasks like expense reports, customer support, coding, and travel booking. These agents require a new kind of interface—one that’s optimized for automated interaction, not human clicks. Manufact is betting big on this transition.
The Model Context Protocol (MCP): A Universal Connector
The foundation of Manufact’s strategy is the Model Context Protocol (MCP), an open standard developed by Anthropic. Before MCP, integrating AI agents with software meant custom connectors for each tool – a tedious, expensive, and unreliable process. MCP solves this by providing a standardized interface, allowing any AI model to connect to any software system seamlessly.
Think of it as the “USB-C of AI.” Just as USB-C standardized device connectivity, MCP standardizes AI access to software. Adoption has been explosive: over 10,000 public MCP servers are now active, with major players like ChatGPT, Cursor, and Google Gemini supporting the protocol.
Manufact’s Three-Pronged Approach
Manufact is building a suite of tools to capitalize on the MCP revolution:
- mcp-use SDK: An open-source library that lets developers integrate AI agents with MCP tools in as few as six lines of code. It supports various large language models, including local ones.
- Built-in Inspector & Testing Suite: A visual debugging tool that allows developers to test MCP servers and tool execution in a sandbox environment.
- Manufact Cloud: A platform for deploying, scaling, and managing MCP servers with near-instant deployment from GitHub – under 60 seconds.
The company’s strategy mirrors Vercel’s success in simplifying front-end web application deployment. Manufact wants to make it just as easy to deploy AI-powered tool integrations.
The Open-Source Advantage
Manufact’s origin story highlights the power of open-source communities. Founders Pietro Zullo and Luigi Pederzani met at a Zurich co-working space and initially built the mcp-use library out of frustration with the complexity of existing solutions. The library went viral, attracting over 5 million downloads and 9,000 GitHub stars before the company even raised funding. This community validation proved a powerful signal to investors.
The Stakes are High: $52 Billion Market by 2030
The AI agents market is projected to grow from $7.84 billion in 2025 to $52.62 billion by 2030. Manufact is positioning itself at the center of this growth, enabling the next generation of AI-powered software interactions.
The Risk of Becoming a “Dumb Database”
The company warns that software companies failing to adapt risk becoming mere “systems of record” – passive data sources for AI agents. If businesses don’t make their products accessible to AI, they could lose control of the user experience and customer relationships.
The Future of Software is Agentic
Manufact’s goal is ambitious: to capture a significant share of every AI tool call on Earth. They aim to be the infrastructure layer for the agentic software revolution, much like Stripe handles global financial transactions.
The company faces challenges: a small team, unproven revenue model, and competition from cloud giants. But Manufact’s focus on open-source development, coupled with the explosive growth of MCP, positions it as a key player in the future of AI-powered software. The transition from building for humans to building for agents is underway, and Manufact is determined to lead the charge.
