By Blockchain News | Created at 2026-09-02 17:46:59 | Updated at 2026-09-02 18:50:20
1 hour ago
Ted Hisokawa Sep 02, 2026 17:29
Runway introduces Dev MCP, a hosted server enabling AI agents to manage models, debug integrations, and streamline generative workflows.
Runway has unveiled Runway Dev MCP, a hosted Model Context Protocol (MCP) server designed to streamline developer workflows for generative media projects. The new tool allows coding agents to research models, configure Model Routers, and debug failed generations—all from within an integrated development environment (IDE).
Runway Dev MCP eliminates the need for developers to run local MCP servers, offering a centralized hosted solution. Agents can now authenticate seamlessly via OAuth and interact directly with a user’s Runway Developer Portal account. Supported functions include selecting and configuring models, managing cost-optimized Model Routers, and retrieving schema data for precise API calls.
What Makes Dev MCP Unique?
Runway Dev MCP extends the capabilities of the previously launched Runway MCP, which was focused on generative output. This new iteration addresses the development process itself, enabling coding agents to handle model selection, integration, and debugging autonomously.
Instead of manually configuring API requests, agents can now use Dev MCP to analyze model capabilities, costs, and schemas before making calls. The integrated Model Router allows optimization for cost, latency, or quality across multiple models, using Runway’s internal metrics to rank responses. Debugging is also automated: agents can identify failure reasons—such as asset size limits or moderation rejections—and rerun corrected requests in real time.
Seamless Integration for Developers
Runway Dev MCP supports popular coding tools like Cursor, Claude Code, and Codex. Developers can connect their accounts via a shared server URL (https://dev.runwayml.com/mcp) and begin leveraging its capabilities almost immediately. For testing, agents can execute a simple whoami command to verify integration with the Developer Portal.
Runway has designed Dev MCP specifically for engineers, product managers, and solo founders using coding agents to manage generative workflows. By centralizing integration activities, the platform reduces friction and increases token efficiency.
Enterprise-Grade Features
Runway has emphasized enterprise-level reliability in Dev MCP’s design. The platform boasts SOC 2 Type II compliance, built-in content moderation, and 99.9% uptime. It also includes spend management tools and visibility into per-model costs, helping companies optimize their operational expenses.
Enterprise users benefit from additional perks such as dedicated account managers and support from Runway’s technical artists. These features make the platform an attractive choice for developers seeking robust tools to scale generative media projects.
Broader Context
Runway Dev MCP builds on a series of recent launches by Runway aimed at simplifying AI-driven content creation. The original Runway MCP, introduced in May 2026, focused on generating media outputs. In July, the company expanded its offerings with Runway Dev, a unified API platform for developers working with image, video, and character models.
The launch of Dev MCP reflects the growing adoption of MCP as an open protocol pattern for connecting AI agents to external tools and APIs. By hosting the server, Runway eliminates the need for developers to manage API keys locally, increasing security and convenience. This move is likely to further solidify its position as a key player in the generative media space.
Getting Started
Developers can begin using Runway Dev MCP by adding it as a custom connector in their preferred coding environment. The hosted server is available at https://dev.runwayml.com/mcp, and access can be approved via OAuth. For more details, visit Runway’s official website.
With the release of Dev MCP, Runway is doubling down on its commitment to ease-of-use and developer empowerment, setting the stage for more efficient and scalable generative media workflows.
Image source: Shutterstock
