Unla, developed by Amoylab, is a free MCP gateway application designed to connect existing APIs and services to AI agents. Built with Go for its core services and ReactJS (TypeScript) for its web frontend, Unla is distributed via Docker images for flexible deployment. It functions as a configurable proxy, transforming various API types into MCP-compliant services. This allows organizations to quickly make their current API functions callable by AI clients without requiring code modifications.

What does Unla connect to?

Unla functions as an MCP Gateway, designed to connect to existing MCP Servers and various APIs, including standard RESTful services. Its core purpose is to transform these connected backend services into MCP-compliant services, making them accessible to AI clients. Operating as a configurable proxy service, Unla allows for the unification of diverse API landscapes under a single, standardized protocol. This architecture ensures that different backend systems can be uniformly presented and interacted with by any compatible MCP Client, centralizing access and management through its interface.

Core features and tools

Unla itself does not expose a predefined set of callable tools. Instead, its primary feature is enabling users to define and expose tools from their *existing* APIs through YAML configuration. This process effectively makes specific API functions callable by AI clients. Unla supports several transport protocols, including MCP SSE, MCP SSE Message, and MCP Streamable HTTP. It is engineered for compatibility with any MCP Client adhering to the MCP Protocol, ensuring broad interoperability for AI agent interactions.

Who is it for?

Unla targets individuals and organizations aiming to quickly make their existing APIs or services accessible to AI agents. It is particularly well-suited for hackathons, prototypes, and smaller-scale projects where rapid setup is essential. This gateway helps convert legacy or RESTful APIs into MCP endpoints for AI agent consumption, providing a unified interface. It simplifies integrating AI capabilities with established backend systems. Skip this if your project does not involve integrating existing APIs with AI agents via the MCP protocol.

How to get started

The server authenticates using environment variables for JWT-based API server access and administrator credentials for the web management UI. Specifically, APISERVER_JWT_SECRET_KEY and SUPER_ADMIN_USERNAME are used, supporting MCP SSE, MCP SSE Message, and MCP Streamable HTTP protocols.

Use cases

• Converting legacy APIs to MCP servers using YAML configuration without code changes, allowing AI agents to interact with them.
• Instantly transforming existing RESTful APIs into MCP endpoints for AI agent consumption.
• Providing a centralized gateway for managing and exposing multiple MCP servers or APIs through a single, unified interface with a built-in management UI.

Unla offers a direct path for integrating existing APIs with AI agents through the MCP protocol. Its Docker-based distribution and YAML configuration for tool exposure simplify deployment and management. The platform is ideal for projects needing quick API transformation and centralized control, especially for hackathons or prototypes. Try converting an existing RESTful API into an MCP endpoint using Unla's YAML configuration.