About Glide
Glide MCP Server connects AI tools to GlideOS for building and operating apps, data, and workflows. Its broad tool surface covers projects, app lifecycle, databases, workflows, files, and backend code.
Key features
- 01Create and manage projects
- 02Publish, promote, and roll back apps
- 03Query and modify app databases
- 04Run workflows and manage triggers
- 05Read, write, and search project files
- 06Run backend code
Use cases
- 01Create and administer GlideOS projects
- 02Inspect and update app database data
- 03Operate scheduled workflows and webhooks
- 04Publish or roll back an app
MCP support
The server exposes GlideOS project, app, data, workflow, file, and code operations to MCP clients over Streamable HTTP. Authentication uses OAuth 2.1 with a Glide account.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://mcp.glideapps.dev/mcpclaude mcp add --transport http dev-glideapps https://mcp.glideapps.dev/mcp{"mcpServers":{"dev-glideapps":{"type":"http","url":"https://mcp.glideapps.dev/mcp"}}}Open in Cursor{"servers":{"dev-glideapps":{"type":"http","url":"https://mcp.glideapps.dev/mcp"}}}Open in VS Code{"mcpServers":{"dev-glideapps":{"type":"http","url":"https://mcp.glideapps.dev/mcp"}}}{"mcpServers":{"dev-glideapps":{"serverUrl":"https://mcp.glideapps.dev/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 READMEGlide MCP Server
Connect your AI tools to Glide's official MCP server and build, manage, and operate GlideOS apps, data, and workflows — from Claude, Cursor, VS Code, Gemini CLI, ChatGPT, or any MCP client.
- Endpoint:
https://mcp.glideapps.dev/mcp(Streamable HTTP) - Authentication: OAuth 2.1 — sign in with your Glide account when your client prompts you; no API keys to manage. Access requires a Glide plan that includes the MCP server feature.
- Documentation: https://www.glideapps.com/docs/os/mcp
What you can do
The server exposes Glide's full MCP tool surface, including:
- Projects — create, list, and manage GlideOS projects
- Apps — publish, promote, roll back, and manage app access and domains
- Data — query and modify app databases, ingest data sources, manage backups
- Workflows — run workflows, manage schedules, webhooks, and triggers
- Files & code — read, write, and search project files; run backend code
Ask your AI tool things like:
"List my Glide projects"
"Create a new project called Inventory Tracker"
"Show me the tables in my app's database"
Setup
Claude Code
claude mcp add --transport http glide https://mcp.glideapps.dev/mcp
Claude (web and desktop)
Go to Settings → Connectors → Add custom connector and enter https://mcp.glideapps.dev/mcp.
Cursor
Install the Glide plugin from the Cursor Marketplace: open Settings → Plugins, search for Glide, and click Install. Or run /add-plugin glide in chat.
The plugin lives in plugins/glide and requires Cursor 3.13.0 or later.
To configure the MCP server manually instead:
Or add to ~/.cursor/mcp.json:
{
"mcpServers": {
"glide": {
"type": "http",
"url": "https://mcp.glideapps.dev/mcp"
}
}
}
Grok Build
The plugins/glide directory doubles as a Grok Build plugin: .grok-plugin/plugin.json, .mcp.json, and a glideos-development skill. Install it from Grok Build's plugin marketplace once the catalog entry is merged — see Listing in the xAI plugin marketplace.
VS Code
code --add-mcp '{"name":"glide","type":"http","url":"https://mcp.glideapps.dev/mcp"}'
Gemini CLI
This repository is also a Gemini CLI extension:
gemini extensions install https://github.com/glideapps/mcp
Then restart Gemini CLI.
ChatGPT
In Settings → Connectors, add a custom connector with the URL https://mcp.glideapps.dev/mcp (requires a ChatGPT plan with connector support).
Repository layout
This repo packages the same MCP server for several agent ecosystems:
| Path | Format | Used by |
|---|---|---|
.cursor-plugin/marketplace.json, plugins/glide/ | Cursor plugin | Cursor Marketplace |
plugin.json, mcp.json | Agent Plugin | Agent Plugin clients |
plugins/glide/.grok-plugin/plugin.json, plugins/glide/.mcp.json | Grok Build plugin | Grok Build (xAI plugin marketplace) |
gemini-extension.json, GEMINI.md | Gemini CLI extension | Gemini CLI |
Validate the Cursor plugin manifests with:
node scripts/validate-template.mjs
Listing in the xAI plugin marketplace
Grok Build installs plugins from xai-org/plugin-marketplace. Listing is one PR to that repo that adds a remote entry pointing at plugins/glide in this repo, pinned to an exact commit:
{
"name": "glide",
"description": "Glide's official MCP server. Build, manage, and operate GlideOS apps: create and manage projects, query and modify app databases, run workflows, publish and roll back apps, and read, write, and search project files.",
"category": "development",
"source": {
"source": "url",
"url": "https://github.com/glideapps/mcp.git",
"sha": "<40-character commit SHA>",
"path": "plugins/glide"
},
"homepage": "https://www.glideapps.com/docs/os/mcp",
"keywords": ["glide", "glideos", "glide apps", "glideapps"],
"domains": ["glideapps.com"]
}
Get the commit to pin (a full 40-character lowercase SHA — branches, tags, and short SHAs are rejected):
git ls-remote https://github.com/glideapps/mcp.git HEAD
Then, in a fork of the marketplace repo, add the entry to .grok-plugin/marketplace.json and run what CI runs before opening the PR:
python3 scripts/generate-plugin-index.py
python3 scripts/validate-catalog.py
python3 scripts/generate-plugin-index.py --check
To ship a plugin update after this repo changes, bump the pinned sha in that entry and regenerate the index — don't open a second entry.
Support
- Glide documentation: https://www.glideapps.com/docs
- Product questions and account help: https://www.glideapps.com/support
- Bugs in the Cursor plugin, Grok Build plugin, Gemini CLI extension, or Agent Plugin manifests, or errors in this README: open an issue on this repository
License
MIT
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.