Self-hosted AWS (AWS Labs) MCP alternatives
The AWS Labs server already self-hosts: it installs locally and runs the AWS CLI through call_aws, with validation, a read-only mode, and suggestions, so the process and your credentials stay on your own machine. If you are reading this cut, you most likely want the same local posture pointed at a different provider, database, or platform.
Every server below installs locally and talks to your agent over stdio. The provider data still travels to each one's own API, but the server process and its tokens stay on infrastructure you control.
The 8 best self-hosted alternatives
- AzureOfficial
The nearest like-for-like: Microsoft's official Azure server installs locally and manages 40-plus services, including storage, Key Vault, Cosmos DB, SQL, and Monitor. The same local CLI-style posture as AWS Labs, on a different cloud.
Set up Azure → - DigitalOceanOfficial
DigitalOcean's official server runs locally and covers App Platform, Droplets, managed databases, Kubernetes, networking, and storage. A simpler provider surface than the full AWS CLI, with the process under your control.
Set up DigitalOcean → - Google Cloud RunOfficial
Narrower than a whole cloud, the official Google Cloud Run server installs locally to deploy local code, list and inspect services, and read logs. Focused on shipping containers from a process you run.
Set up Google Cloud Run → - HerokuOfficial
Heroku's official Platform server runs locally to deploy apps, scale dynos, manage add-ons and pipelines, run one-off commands, and operate Postgres. An app-platform model rather than raw cloud, kept on your machine.
Set up Heroku → - NetlifyOfficial
Front-end deploys are the focus, and the official Netlify server installs locally to create, deploy, and manage projects, env vars, forms, access controls, and extensions. The local pick when full infrastructure is not the job.
Set up Netlify → Observability rather than provisioning: Grafana Labs' official server runs locally and queries dashboards, Prometheus, Loki, incidents, alerts, and OnCall. Adjacent to AWS, it fits watching what you have deployed.
Set up Grafana →For a local database rather than a cloud, the DBHub SQLite server runs an agent against a SQLite file with execute_sql and search_objects. Zero-dependency and entirely on your disk, well outside AWS's provisioning scope.
Set up SQLite (DBHub) →A DSN connects an agent to a database through the DBHub Postgres gateway, with execute_sql and search_objects, all from a local process. Reach for it when the job is querying your own Postgres, not managing AWS services.
Set up DBHub (Postgres) →
How to choose
For a self-hosted full-cloud server like AWS Labs, Azure is the closest, with DigitalOcean a simpler provider and GCP, Heroku, and Netlify narrower deploy targets, all running locally. Grafana, SQLite, and Postgres are adjacent: observability and databases rather than provisioning. One caveat: self-hosting controls where the process and credentials live, but each provider's data still travels to its own API.
FAQ
- Is the AWS MCP server self-hosted?
- Yes. The AWS Labs server installs locally and runs the AWS CLI on your own machine, with validation and a read-only mode, so the process and your credentials stay with you. Every alternative on this page also runs locally over stdio.
- Does running the server locally keep my cloud data on my own machine?
- It keeps the MCP server process and its credentials on your infrastructure, which is usually the point. The provider data itself still goes to each service's API, whether that is Azure, DigitalOcean, or Heroku. The exceptions here are SQLite and Postgres, which query a database you already control.