- Published on
Rayrun β one MCP gateway for the whole company
- Authors
- Name
- DeepHorus
Rayrun has launched a new platform designed to streamline the deployment and management of Model Context Protocol (MCP) servers for coding agents, offering a unified gateway for entire companies. This service allows developers to run multiple MCP tools under a single URL while managing credentials and access controls centrally, simplifying a process that often requires fragmented client configurations. The launch, announced on Hacker News, provides a free tier for individual developers without requiring a credit card, aiming to lower the barrier to entry for AI-powered tool integration.
The platform emerges as coding agents like Codex and Claude Code become more prevalent in software development workflows. Traditionally, each client or tool needed its own server and configuration, leading to operational complexity. Rayrun addresses this by having the agent build the server, then deploying it behind a single endpoint with centralized credential management. This approach promises to replace the need for multiple client setups with one streamlined service.
Key features include a one-command deployment process using rayrun deploy, which streams build and release status directly to the terminal. The platform supports multiple deployment sources: remote MCP endpoints (with OAuth completion), OpenAPI specifications (auto-converted to typed tools), pinned npm packages, and container images. Access control is granular, decided per service, tool, and argument, with an option for human approval on certain calls.
Security and compliance are highlighted as core components for enterprise teams. Rayrun employs workspace-specific encryption, ensures secrets never reach browsers, and supports single sign-on (SAML 2.0/OIDC) with SCIM provisioning. Every API call generates a signed, searchable audit record that can be streamed to a SIEM. The platform's controls and third-party data-sharing policies are published transparently for security reviews.
The service is positioned to reduce the operational overhead for teams integrating AI agents into their development pipelines. By handling tool deployment, credential storage, and access policies in one place, it allows developers to focus on building the tool's logic rather than its infrastructure. The initial offering starts with a single server, encouraging teams to connect services incrementally as needs grow.
