In addition to the existing TypeScript or JavaScript based managed MCP servers, Metorial now supports Python-based ones. This one is for the Python lovers :)

Businesses and enterprises can now connect their SAML or OAuth based authentication services, like Okta, Entra, and Rippling, to Metorial. Metorial's authentication infra then passes along the authentication state to our MCP platform to enable true single sign on for MCP.

Until now Metorial was focused on API-controlled access to MCP servers. Today we're invading Cursor, Claude Code, and more with Magic MCP. Magic MCP is a single URL you can use to connect all of your agents and vibe coding tools to our reliable platform.

MCP servers don't support webhooks from sources like Stripe or Github. From now on, they do using Metorial Callbacks. Callbacks allow MCP servers to receive webhooks from external providers and the forward them to your agents to process them. It's as easy as one line of code and we handle the rest, including monitoring, retires, and more.

Yup, we just open sourced Metorial. Read the code, contribute, or get inspired. Check it out at github.com/metorial/metorial-platform

Starting today, anyone can publish MCP servers on Metorial. Rest assured that these servers are vetted and verified by the Metorial team. This now enabled teams building their own MCP servers on Metorial to share them with the community. We are already working with select MCP server providers to publish their servers on our platform.

Enterprise deployments need advanced security for MCP servers. Metorial already isolates every MCP connection, connecting it to it's own MCP server isolated from others to avoid problems like token leakage. Today we are taking this one step further by allowing you to restrict where an MCP server can connect to so you can be assured it only connects where it is supposed to.

With managed servers you can now deploy your own MCP servers on Metorial. These servers run on the same super reliable platform that already powers our built-in MCP servers. Get started by creating a new managed server in the Metorial dashboard - it deploys in seconds and scales to thousands of connections with monitoring and our great SDKs on top.

With oauth connections you can now set up authentication to MCP servers on Metorial using OAuth. This works with any MCP server - not matter if it natively supports OAuth or now. We handle all of the complicated parts like authentication errors, token refresh, and security.

Starting today, every MCP server running on Metorial is running on our new custom MCP engine. Think of this as our own Kubernetes built just for MCP. It's the scaling and orchestration layer for serverless MCP. What this means for you is that MCP connections to Metorial are now more reliable, faster, and you can have tens of thousands of them.

Providers like Linear or Neon offer their own remote MCP servers. Metorial now has the ability to connect these providers just like you would connect any other MCP server and you get our SDKs, monitoring, and reliable platform on top.

Metorial's new Python and TypeScript SDKs enable you to connect the Metorial Integration Platform to the AI Agent SDKs you are already using. We have support for Vercel's AI SDK, the OpenAI, Anthropic, and Google SDKs, and even LangChain.

Check out our examples on GitHub:

• Metorial TypeScript
• Metorial Python

Connect to any integration hosted on Metorial using our serverless gateway. When you connect, we boot up an MCP server in less than a second. Connections can run for day or even months while only incurring compute costs when they are needed. You can even reconnect to MCP connections weeks of months after they have been started. Of course all of this is integrated with Metorial APIs and SDKs to you can connect from JS or Python in just one line of code.

The Metorial Integration Platform makes it easy to connect any AI agents to external tools and data sources through MCP. All of this is enabled by our serverless platform specifically designed for MCP. The Metorial platform can run any MCP server with sub-second cold starts. Things like monitoring, security, and scaling are handled automatically.

We’ve introduced Model Context Protocol (MCP) Containers—a new way to run and manage AI model servers more efficiently. Over 400 MCP servers have now been fully containerized, making it easier to deploy, scale, and maintain AI infrastructure.

The best part: it’s completely open source. You can explore the code, contribute, or start using MCP Containers in your own workflows.

Learn more: https://github.com/metorial/mcp-containers

We’re excited to roll out Chat Message Streaming, bringing faster and more dynamic AI interactions to Metorial Portals and the Metorial API!

• Real-time message streaming – AI responses are streamed as they are generated, reducing wait times.
• Seamless Portal integration – Users see messages appear instantly, just like ChatGPT.
• API support – Developers can now receive streamed responses for a smoother, more interactive experience.

This update makes AI conversations feel more natural, fluid, and engaging—whether you’re using Metorial Portals or integrating via the API.

We’ve given the Metorial Dashboard a major upgrade, making it sleeker, faster, and more user-friendly than ever!

• New Sidebar Layout – Navigate effortlessly with an improved, streamlined sidebar.
• Integrated Documentation – Access Metorial Docs directly from the dashboard for a seamless workflow.
• Built-in Feedback & Support – Get help and share feedback without leaving the platform.
• Faster Load Times – Enjoy a smoother, more responsive experience.
• A Stunning New UI – A fresh, modern look that’s as functional as it is beautiful.

🚀 Check out the new dashboard and experience the upgrade today!

We’re thrilled to launch Metorial Portals, a powerful way to create custom ChatGPT-like websites connected to your AI agents—no coding required!

With Portals, you can:

• Create standalone or embeddable AI chat experiences tailored to your needs.
• Fully customize the look and feel to match your brand.
• Define agent behavior to control how your AI interacts with users.
• Seamlessly integrate user authentication for secure and personalized conversations.
• Leverage a beautifully designed UI for an intuitive and engaging chat experience.

Whether you're building an AI-powered customer support assistant, knowledge base, or interactive chatbot, Metorial Portals make it effortless to deploy.

🚀 Start building your AI-powered website today!

We’re excited to introduce support for image models like DALL·E, enabling powerful AI-driven image generation directly within Metorial!

With this update, you can:

• Generate images using AI.
• Combine image models with language models to create rich, multimodal experiences.
• Seamlessly integrate AI-generated visuals into your applications and workflows.

Whether you're building creative tools, automated design assistants, or AI-powered storytelling experiences, Metorial now makes it easier than ever to blend text and image generation for limitless possibilities.

Start creating with AI-powered images today!

We’re excited to launch the Metorial API, giving developers full programmatic control over their AI Agents! Now, you can seamlessly integrate AI-powered interactions into your applications, products, and workflows.

With the Metorial API, you can:

• Manage conversation threads – Create, update, and retrieve conversations effortlessly.
• Send and receive messages – Programmatically interact with your AI Agents in real time.
• Automate workflows – Build custom AI-driven applications with ease.
• Scale AI interactions – Deploy AI capabilities across multiple use cases, from chatbots to automation tools.

Whether you're building a custom AI-powered app, automating customer support, or integrating AI into existing software, the Metorial API gives you unmatched flexibility and control.

Get started today – Check out our API documentation and start building! 🚀

New Features

• Pre-Built Agent Templates – Get started faster with ready-to-use agent templates.
• Pro Plans – Unlock advanced features with various Pro plan options.

Improvements

• Multi-Project Support – Manage multiple projects seamlessly.
• Team Collaboration – Invite team members and collaborate effortlessly.

New Features

• Google Gemini Model Support – Now compatible with Google Gemini models.
• OpenAI GPT Model Support – Added support for OpenAI GPT models.
• Custom Code Tools – Enhance agents with custom JavaScript code for greater flexibility.

Improvements

• Agent Environments – Introduced sandboxed environments for better agent isolation.
• Agent Test Runs – Users can now test their agents directly in the Metorial Dashboard.
• Enhanced Logging – Added logs for users, runs, and threads to improve tracking and debugging.
• Detailed Pages – New detail views for users, runs, and threads for better insights.

Bug Fixes

• Various bug fixes and minor improvements for a smoother experience.

With the rapid advancements in AI, tools like OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini are more powerful than ever. However, creating complex AI interactions remains a challenge—until now.

Metorial is a high-level visual AI builder that simplifies the process with an intuitive drag-and-drop interface. Whether you're integrating pre-trained models or crafting custom AI agents, Metorial makes AI development accessible to everyone.

Seamless code and no-code integrations ensure that AI-powered tools can be added to any app or website in minutes. Whether you're a developer or a no-code enthusiast, Metorial bridges the gap between APIs and visual AI workflows—empowering innovation without limits.

Be among the first to experience the next generation of AI development. Sign up for updates today!