Mcp Nhost

Overview

MCP‑Nhost serves as a bridge between AI assistants and Nhost Cloud, enabling developers to manage projects, run queries, and modify configurations directly from a conversational interface. By exposing Nhost’s GraphQL endpoints through the Model Context Protocol, it eliminates the need for manual API calls or CLI interactions. Developers can ask an AI assistant to retrieve a project’s schema, execute a mutation that updates permissions, or fetch documentation links—all within the same chat flow. This streamlines workflows and reduces context switching between tools.

The server offers a comprehensive set of tools that cover both cloud‑hosted and local development environments. For the cloud, it provides and , allowing AI assistants to introspect the platform’s schema and perform operations such as creating projects or managing organizations. Local tools mirror this functionality, giving developers the ability to query local Hasura instances () and manage migrations through . The inclusion of configuration tools ( and ) lets assistants modify runtime settings, ensuring that changes are reflected after a simple . This breadth of capabilities means the same MCP server can support end‑to‑end development, from schema discovery to deployment.

Key features include:

• Unified GraphQL access for both cloud and local instances, with role‑based execution defaults (e.g., “user” role).
• Schema introspection tools that expose available queries, mutations, and types, helping assistants generate accurate prompts or code snippets.
• Management operations for Hasura metadata, migrations, and permissions via , enabling AI‑driven database evolution.
• Documentation search through the tool, which returns direct links to Nhost guides and API references.
• Mutation control—mutations must be explicitly enabled when starting the server, providing security and preventing accidental changes.

Real‑world scenarios include a developer asking an AI assistant to “create a new Hasura migration that adds a column to the table” and having the assistant run to apply it. In another case, a product manager might query the cloud schema to confirm that a new feature’s GraphQL endpoint is available, or retrieve documentation links for onboarding new team members. Because MCP‑Nhost exposes all these operations through a single protocol, it fits naturally into AI‑augmented workflows where assistants act as both helpers and orchestrators.

What sets MCP‑Nhost apart is its tight coupling with Nhost’s own tooling. By leveraging the same GraphQL schemas that Nhost CLI and dashboard use, it ensures consistency across environments. The ability to toggle mutation permissions at startup gives teams granular control over what an AI assistant can modify, striking a balance between automation and safety. Overall, MCP‑Nhost turns the Nhost ecosystem into an interactive, AI‑friendly service that accelerates development and reduces friction in managing cloud projects.