About Atlas Cloud
Atlas Cloud is an MCP integration for deploying and managing compatible web apps through an AI coding assistant. It connects to Atlas hosting, isolated Postgres, and project file storage, with tools for deployment and project inspection.
Key features
- 01Deploy private source uploads
- 02Redeploy saved source
- 03Configure environment settings
- 04Inspect isolated databases and run migrations
- 05Retrieve files and inspect logs
- 06View reported usage
Use cases
- 01Deploy a web app with Postgres and file storage
- 02Inspect a build and recent runtime errors
- 03Update configuration and redeploy a project
MCP support
Connects MCP clients to Atlas Cloud through a remote Streamable HTTP endpoint. OAuth with Dynamic Client Registration grants approved workspace read and deployment access.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://mcp.atlasos.one/api/mcpclaude mcp add --transport http brandworks-technologies-atlas-cloud https://mcp.atlasos.one/api/mcp{"mcpServers":{"brandworks-technologies-atlas-cloud":{"type":"http","url":"https://mcp.atlasos.one/api/mcp"}}}Open in Cursor{"servers":{"brandworks-technologies-atlas-cloud":{"type":"http","url":"https://mcp.atlasos.one/api/mcp"}}}Open in VS Code{"mcpServers":{"brandworks-technologies-atlas-cloud":{"type":"http","url":"https://mcp.atlasos.one/api/mcp"}}}{"mcpServers":{"brandworks-technologies-atlas-cloud":{"serverUrl":"https://mcp.atlasos.one/api/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMEAtlas Cloud
![]()
Deploy and manage compatible web apps with Atlas hosting, isolated Postgres and project file storage through your AI coding assistant.
This is the public integration package. It contains an opt-in deployment skill, client configuration, registry metadata and a small example. Atlas's backend, infrastructure credentials, customer data and reviewer credentials are not included. This repository has independent history and contains no backend source.
Connect
Remote MCP endpoint: https://mcp.atlasos.one/api/mcp
Transport: Streamable HTTP. Authentication: OAuth with Dynamic Client Registration. Users sign in to Atlas, select their workspace and approve read/deployment access. No shared API key is packaged.
An Atlas account with verified email, workspace permissions and billing enrollment is required for deployment. Hosting and other infrastructure usage can incur Atlas service charges; see Billing. A coding session needs local source, archive creation and HTTPS upload access for direct source deployment.
Cursor
Add the following server using Cursor's MCP settings or merge it into your existing MCP configuration. Preserve unrelated entries. Authenticate through the browser flow offered by Cursor.
{
"mcpServers": {
"atlas": { "url": "https://mcp.atlasos.one/api/mcp" }
}
}
Codex
codex mcp add atlas --url https://mcp.atlasos.one/api/mcp
codex mcp login atlas --scopes atlas:read,atlas:deploy
Install the optional skill by copying only skills/atlas-deploy into your user skills directory. Connection details.
Claude Code
claude mcp add --transport http atlas https://mcp.atlasos.one/api/mcp
Open /mcp and authenticate Atlas. Alternatively, load this package locally with claude --plugin-dir . and authenticate its bundled server; avoid adding a duplicate server.
Example requests
- Deploy this app with Atlas Cloud, an isolated Postgres database and file storage. Verify the live URL.
- Inspect this Atlas project's latest build and runtime errors without making changes.
- Update this project's configuration, redeploy the same project once, and verify the app.
The optional deployment skill applies when users choose Atlas or their project already uses Atlas. It does not override another hosting choice or authorize changes by itself.
Capabilities and limits
- Direct private source upload and updates to an existing project. Public repository publication is not required.
- Saved-source redeployment, write-only custom environment settings, and isolated database inspection/migrations within approved access.
- Bounded file retrieval, build logs, recent runtime logs and reported usage.
- Next.js and Vite selectors; other web runtimes require compatibility checks. Dedicated GPU workers and arbitrary persistent processes are unsupported.
- Source archive: 10 MiB compressed and total file contents, 2 MiB per file, 1,500 entries. Runtime log queries cover a recent 30-minute window; reporting can lag. Missing usage is unavailable, not zero.
- Destructive actions can change application behavior or database data. OAuth workspace permissions and runtime enforcement remain in the hosted service.
Example
examples/atlas-hello is a static Vite greeting app with no credentials or database. It is a first-party example, not an independent customer case study.
Documentation and support
Guides · Privacy · Terms · Status
Support and security reports: [email protected]. Keep secrets and customer data out of public issues. Connected assistants receive the scoped tool inputs/results needed for the requested operations.
Registry listing
Atlas Cloud metadata is published in the Official MCP Registry as io.github.Brandworks-Technologies/atlas-cloud. A registry listing is not an endorsement, security certification or automatic installation. OAuth approval is still required.
License and service
MIT covers only the files in this repository. Atlas branding and the hosted service are not licensed as backend source; hosted access is governed by Atlas's service terms. Publishing this package does not establish directory approval, endorsement, automatic installation or a free hosting entitlement.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.