ACI MCP Servers

The Aipolabs MCP server package delivers two distinct Model Context Protocol (MCP) services that let AI assistants tap into the full power of ACI.dev’s managed functions. The first service, aci‑mcp‑apps, exposes a curated set of tools from one or more specified ACI.dev applications. This is ideal when an assistant needs to work within a constrained domain—such as only searching the web or sending emails—while still leveraging the robust, cloud‑hosted logic that ACI.dev provides. The second service, aci‑mcp‑unified, offers a meta‑layer that dynamically discovers every function available on the ACI.dev platform and lets an assistant invoke any of them through two simple commands: and . This “universal” gateway removes the need for manual registration of each tool, enabling rapid experimentation and deployment across a wide range of use cases.

For developers building AI workflows, these servers solve the problem of tight coupling between an assistant’s code and its toolset. By running as independent MCP services, they decouple the logic that implements business rules from the assistant’s core, allowing teams to update or add functions on ACI.dev without touching the assistant code. This promotes a clean separation of concerns, easier maintenance, and faster iteration cycles. Moreover, because the servers are lightweight Python packages that can be launched via or Docker, they fit seamlessly into CI/CD pipelines and containerized environments.

Key capabilities include:

• Dynamic tool discovery with , giving assistants an up‑to‑date inventory of available actions.
• Safe, authenticated execution via , ensuring that only authorized calls reach the underlying services.
• App‑specific filtering in , which limits tool exposure to a selected set of applications, enhancing security and compliance.
• Built‑in debugging support through the MCP inspector and log forwarding, making it straightforward to trace interactions between assistants and tools.

Typical use cases span from internal productivity bots that need to read emails, schedule meetings, or query databases, to public-facing assistants that pull in weather data, stock prices, or custom business logic. By routing these actions through ACI.dev’s managed functions, organizations can centralize governance, monitor usage, and enforce rate limits—all while keeping the assistant’s codebase lean. The result is a scalable, maintainable AI ecosystem where new tools can be added or retired with minimal friction, and developers can focus on crafting richer conversational experiences rather than plumbing infrastructure.