Manufact Raises $6.3 Million to Build Infrastructure for MCP Developers

SAN FRANCISCO, Calif. - Manufact has raised $6.3 million in seed financing to expand its open-source tools and cloud infrastructure for Model Context Protocol servers, marking a significant bet on the niche but critical plumbing required for agentic artificial intelligence. Peak XV Partners led the round, with participation from Liquid 2 Ventures, Ritual Capital, Pioneer Fund, Y Combinator, and other backers. The investment arrives as developers increasingly seek standardized methods for connecting large language models to external data sources and business applications, a task that has historically required bespoke, brittle integrations.

The San Francisco startup was previously called mcp-use, the name of its developer project that gained early traction within the open-source community. This rebranding represents a shift from a community-driven experiment to a commercial entity focused on enterprise-grade reliability. The funding will allow the small team to deepen its platform capabilities and support more complex production workloads, transitioning from a toolkit for hobbyists to a managed layer for professional software engineering teams.

At its core, Manufact's software helps teams build, test, and host Model Context Protocol (MCP) servers. These servers act as the bridge between AI assistants and the outside world, providing standardized access to diverse applications and datasets. Without such a protocol, every new integration between an AI model and a database or a SaaS tool would require unique code, creating a maintenance burden that scales poorly as AI products become more integrated into existing business workflows.

Manufact is betting that developers will prefer a managed layer as these connections move into production environments. While building a single prototype connector is relatively straightforward, maintaining dozens of them in a live setting introduces significant overhead. A managed platform can handle the underlying infrastructure, allowing product teams to focus on the logic of their AI applications rather than the logistics of server uptime and connectivity.

The Model Context Protocol has rapidly become a common way to connect models with business systems, gaining momentum as organizations look beyond simple chat interfaces toward autonomous agents. However, the emerging ecosystem remains highly fragmented. Dozens of developers are creating their own MCP servers for various tools, but there is little consistency in how these connectors are deployed or monitored. Manufact views this fragmentation as an opportunity to provide a unifying layer.

A hosted service can remove the heavy lifting of deployment and observability from product teams while preserving the option to use open-source components. This hybrid approach is a staple of modern developer tools, allowing users to start with free, transparent code while eventually paying for the convenience and security of a managed cloud environment. By offering a hosted path, Manufact seeks to capture the value generated by the open-source project it originated.

That positioning puts Manufact in a competitive space between individual developers and larger cloud platforms that can offer similar tooling. Giants like Amazon Web Services or Microsoft Azure possess the scale to integrate these protocols into their existing suites, meaning Manufact must offer a superior developer experience or specialized features that generic cloud providers lack. The startup’s success hinges on being more agile and focused on the specific needs of the MCP ecosystem.

The company still has to prove that growing interest in MCP becomes durable revenue. History in the software sector is full of popular protocols that failed to generate significant commercial value for their creators. While developers may love the ease of use provided by Manufact, the transition from an interesting utility to a mission-critical line item in a corporate budget is a difficult one that requires demonstrating clear return on investment.

Furthermore, protocol support alone is unlikely to be a lasting competitive advantage. As MCP becomes more widely adopted, it could become commoditized, with multiple vendors offering similar connectivity features. In such a scenario, Manufact will need to differentiate itself through advanced features that go beyond basic connectivity, such as specialized debugging tools, performance optimizations, or unique data transformation capabilities.

Enterprise buyers will expect more than just connectivity; they will demand strong security, granular permissions, and high reliability. For an AI to safely access internal databases or customer information, the connectors must be hardened against vulnerabilities. Manufact will need to invest heavily in its security posture to satisfy the compliance requirements of large-scale organizations that are wary of the risks associated with AI integrations.

Manufact will also need to walk a fine line between keeping its open-source community engaged while charging for the managed services that fund the business. If the company moves too many features behind a paywall, it risks alienating the developers who helped build the protocol's momentum. Conversely, if the managed service does not offer enough unique value, there will be little incentive for users to pay for what they can run themselves on-premise.

The seed round gives the Manufact team the necessary capital to scale their engineering efforts and build out the features required for enterprise adoption. The next test for the company is whether customers use Manufact for ongoing, high-stakes operations, rather than only during early-stage experiments. Successfully moving clients from the pilot phase to the production stage will be the primary metric for the company's long-term viability.

Industry analysts have noted that the market for AI infrastructure is currently seeing a surge in funding, but the winners will be those who can provide stability in a volatile technical landscape. As the underlying models from companies like OpenAI and Anthropic evolve, the tools that connect them to the real world must remain compatible and performant, placing constant pressure on infrastructure providers like Manufact to stay at the cutting edge.

If it can become a trusted deployment layer, the company could benefit significantly as AI products rely on an ever-increasing number of external tools. The shift toward agentic AI, where models perform actions rather than just generating text, makes robust connectors even more essential. In this future, the infrastructure that manages the dialogue between the model and the database becomes the central nervous system of the enterprise AI stack.

For now, Manufact remains a small team with a large ambition: to define how AI interacts with the digital world. The success of this $6.3 million seed round reflects the venture capital community's belief that the infrastructure for AI connectivity is just as important as the models themselves. As the company moves forward, its ability to execute on its technical roadmap while building a sustainable business model will be closely watched by the industry.