mcpField
S
Steep

Steep

Remote
steep.appData & Analytics9Updated 10w ago

About Steep

Steep MCP connects AI agents to a Steep workspace for read-only analytics queries. It is for teams using MCP-compatible clients to inspect metrics, targets, entities, modules, members, and teams through natural language.

Key features

  1. 01List workspace metrics
  2. 02Query metric data by date range
  3. 03Slice and break down metric results
  4. 04List targets and query progress
  5. 05Explore entities and their modules
  6. 06List workspace members and teams

Use cases

  1. 01Compare metric results across a selected date range
  2. 02Check progress toward a team target
  3. 03Find entities that can slice a metric
  4. 04Identify members of a workspace team

MCP support

Steep exposes a hosted remote MCP server at `https://mcp.steep.app/mcp` using Streamable HTTP. Its tools are read-only and clients authenticate with OAuth.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://mcp.steep.app/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

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 README

Steep MCP

This repository contains the configuration needed to integrate Steep with Claude Code, Cursor, and other MCP-compatible clients. The plugin lets your AI agents interact directly with your Steep workspace, querying metrics, targets, entities, and team data through natural language.

Features

The Steep MCP server provides the following capabilities:

  • Metrics: list metrics in the workspace and query metric data over any date range, with optional slicing and breakdowns
  • Targets: list targets and query target progress
  • Entities & modules: list entities and the modules they appear in
  • Workspace: list workspace members and teams

All tools are read-only.

Prerequisites

Before setting up the Steep MCP server, ensure you have:

  • Claude Code CLI, Cursor IDE, or another MCP-compatible client installed
  • A Steep account with access to the workspace you want to query

Installation

Claude Code & Cowork

Install from the community plugin marketplace:

/plugin marketplace add anthropics/claude-plugins-community
/plugin install steep@claude-community

Alternatively, install directly from this repo:

git clone https://github.com/SteepHq/steep-mcp.git
claude --plugin-dir ./steep-mcp

Either path loads .claude-plugin/plugin.json and the sibling .mcp.json, configures the Steep MCP server, and prompts you to authenticate via OAuth. No client ID is needed — Claude Code discovers the auth server and registers itself dynamically.

Cursor

Install with one click from the Cursor plugin marketplace.

Alternatively, add Steep manually: Cursor → Settings → MCP and paste:

{
  "mcpServers": {
    "steep": {
      "url": "https://mcp.steep.app/mcp"
    }
  }
}

Save the configuration, then click the connect button to authenticate via OAuth.

Other MCP clients

Point your client at https://mcp.steep.app/mcp (Streamable HTTP transport). Clients supporting Dynamic Client Registration (RFC 7591) will register automatically.

Usage Examples

Once configured, you can interact with Steep through your AI assistant using natural language:

  • List metrics: "What metrics do we have in Steep?"
  • Query data: "Show MRR for the last 6 months, broken down by plan"
  • Track targets: "How are we tracking against our Q2 revenue target?"
  • Explore the data model: "Which entities slice the customer count metric?"
  • Find people: "Who's on the growth team in Steep?"

Updates

Marketplace installs (Claude Code, Cursor) auto-update when we publish a new version. Run /plugin update steep in Claude Code to pull the latest.

Cursor handles updates through its own marketplace UI. Users who installed manually via --plugin-dir or by editing mcp.json directly will keep using whatever revision they cloned — re-clone or re-paste to get changes.

Documentation & resources

Notes & limitations

  • Remote server only: this configuration connects to Steep's hosted MCP server. No local installation is required or supported.
  • Workspace scope: you can only access data your Steep user account already has access to. OAuth scopes are limited to the read scopes the server advertises at /.well-known/oauth-protected-resource.

Questions or Issues?

For product questions and integration help, contact [email protected].

For security reports, see SECURITY.md.

License

Apache-2.0 — see LICENSE.

The Steep name and logo are trademarks of Steep. See NOTICE.

Comments

Markdown supported

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.