MCP Demo Day: How 10 Leading AI Companies Built MCP Servers on Cloudflare
Today, Anthropic, Asana, Atlassian, Block, Intercom, Linear, PayPal, Sentry, Stripe, and Webflow have introduced a new wave of remote Model Context Protocol (MCP) servers built on Cloudflare. These servers empower Claude users to manage projects, generate invoices, query databases, and deploy full-stack applications—all within the chat interface.
Since Anthropic introduced MCP in November, the concept has gained momentum. MCP bridges the gap between AI agents and the tools they use to perform actions and access context. Previously, users had to install MCP servers locally, but with Anthropic’s Integrations, MCP servers are now accessible like any website—just enter a URL and start.
Cloudflare has focused on creating tools that simplify building remote MCP servers. This lets engineering teams concentrate on developing MCP tools for their apps instead of dealing with protocol challenges. Deploying a remote MCP server on Cloudflare takes just one click, with pre-built support for the latest MCP standards.
Leading companies leverage Cloudflare’s platform to deliver AI-powered experiences quickly and reliably. You can join them by clicking “Deploy to Cloudflare” and start building your own MCP server. Read on to learn more about the capabilities these companies offer and how they made it happen. You can also see these solutions in action during Demo Day on May 1 at 10:00 AM PST.
MCP: Powering the Next Generation of Applications
Just like websites became essential for businesses as internet access grew, AI tools are now becoming integral to everyday tasks. Developers often turn to AI tools like Claude to write code, so it makes sense for AI to also help deploy that code. Similarly, AI can assist non-developers with tasks like ordering groceries based on a recipe it helped create.
Remote MCP servers make these scenarios possible. Early adopters of MCP will likely benefit from a first-mover advantage. Faster access to your product’s value means higher chances of user success, upgrades, and retention. By connecting services to AI agents through MCP, users can request what they need and let the agent handle the details—enabling quicker “aha” moments.
Making Your Service AI-First with MCP
- Lower the barrier to entry: Users don’t need to learn complex dashboards or read documentation. They simply describe their needs, and the AI agent figures out the rest.
- Create personalized experiences: MCP tracks user requests and interactions, adapting future tool calls to usage patterns and preferences for relevant, customized experiences.
- Drive upgrades naturally: AI agents can explain how higher-tier plans help users meet their goals instead of relying on feature comparison tables.
- Ship new features and integrations: Expose your tools via MCP to let users connect your product with their existing stack. Agents can combine tools across providers without you handling every third-party integration.
Why Build MCP on Cloudflare?
Cloudflare has made it easier for developers to build and deploy remote MCP servers by handling protocol complexity, authentication, and transport. Cloudflare supports fully stateful MCP servers that “sleep” when idle to reduce costs.
Recent updates include:
- Streamable HTTP transport support: MCP servers can now use the latest transport alongside SSE. Cloudflare’s AI Playground and mcp-remote proxy support this for easy testing.
- Python support: Build MCP servers using Python in addition to JavaScript/TypeScript.
- Improved docs and templates: Quickstart templates, expanded documentation, and best practices based on real-world experience.
Here’s what some companies have built and why they chose Cloudflare for their MCP deployment.
Asana
Asana uses MCP to let users interact with its Work Graph through natural language. Tasks like getting project updates, managing deadlines, and sending comments can be done from any MCP client. This enables smooth integration across tools and workflows, such as converting meeting notes into assigned tasks or pulling Asana tasks into MCP-enabled IDEs.
Asana chose Cloudflare’s infrastructure to accelerate deployment and focus on user experience while Cloudflare handled authentication and scaling.
“MCP connects our Work Graph directly to AI tools like Claude.ai, enabling AI to become a true teammate in work management.” – Prashant Pandey, CTO, Asana
Atlassian
Jira and Confluence Cloud customers can now interact with their data securely from Claude using Atlassian’s Remote MCP Server, hosted on Cloudflare. Users can summarize work items, create issues or pages, perform multi-step actions, and enrich Jira items with context—all while keeping data secure.
Cloudflare’s platform provided OAuth, remote MCP support, and scalability, enabling Atlassian to build and deploy quickly.
“Hosting on Cloudflare infrastructure means we can bring this integration to customers faster while keeping enterprise data secure.” – Taroon Mandhana, Head of Product Engineering, Atlassian
Intercom
Intercom’s AI agent Fin autonomously resolves over 50% of customer support conversations. The Intercom MCP server makes customer conversation data accessible via AI tools, helping teams diagnose issues and improve service efficiency.
Cloudflare’s tools helped Intercom launch their MCP server in under a day, simplifying the connection between AI assistants and internal systems.
“Cloudflare's platform helped Intercom go from concept to production quickly, making MCP server launch effortless.” – Jordan Neill, SVP Engineering, Intercom
Linear
Linear’s MCP server brings issue tracking and product workflows into AI assistants. Teams can refine specs, debug, or create issues directly from conversations without context switching.
Building on Cloudflare allowed Linear to leverage fast, flexible edge compute and frameworks to keep users focused and productive across tools.
“MCP helps developers and product teams access their work where they already are, eliminating context switching.” – Tom Moor, Head of US Engineering, Linear
PayPal
PayPal’s remote MCP server lets developers delegate commerce workflows to AI agents using natural language. Tasks like managing inventory, processing payments, tracking shipping, and handling refunds can be executed autonomously.
Developers can start integrating with PayPal’s MCP server on Cloudflare today.
“This is a revolutionary development for commerce.” – Prakhar Mehrotra, SVP of Artificial Intelligence, PayPal
Sentry
Sentry’s MCP server enables developers to query error and issue data directly from their IDE or AI assistants. It supports creating projects, capturing setup details, and will soon include root cause analysis and instrumentation features.
Hosting on Cloudflare sidestepped local scaling and authentication issues. Durable Object support preserves state, which is crucial for non-local MCP servers.
“MCP integrates crucial Sentry context into the developer workflow, empowering teams to build reliable applications.” – David Cramer, CPO and Co-Founder, Sentry
Block (Square)
Square’s MCP server enables sellers to use AI agents to manage payments, orders, inventory, and customer data via Square’s APIs. This lowers technical barriers and allows sophisticated use cases.
Stripe
Stripe views MCP as an emerging AI interface that could become the default way people and code interact with services. Their agent toolkit, combined with Cloudflare’s SDK, makes monetizing MCPs straightforward.
“Developers can monetize their MCPs with just a few lines of code.” – Jeff Weinstein, Product Lead, Stripe
Webflow
Webflow’s MCP server supports CMS management, SEO auditing, content localization, and site publishing through AI agents. This allows users to manage and improve their websites directly from AI tools.
Webflow sees MCP as more than an API surface—it exposes meaningful actions in a secure, contextual way, enabling new workflows where AI and humans collaborate smoothly.
“Cloudflare supports our need for reliability, performance, and developer tooling to build modern web infrastructure.” – Utkarsh Sengar, VP of Engineering, Webflow
Start Building Today
If you want to build a remote MCP server for your service, start with the documentation and tutorials available. You can deploy a starter MCP server to production with one click and then connect it to Claude, Cloudflare’s AI playground, or any remote MCP client.
There’s also a new video walkthrough showing how to build MCP servers using Cloudflare’s templates. For questions or feedback, you can reach out via email at 1800-mcp@cloudflare.com.
For those interested in expanding their AI skills, consider exploring AI courses tailored for product developers at Complete AI Training.
Your membership also unlocks:
