prompeteer-mcp

Prompeteer MCP is an MCP (Model Context Protocol) server from Prompeteer.ai, used to generate contextual prompts and agent skills for more than 140 AI platforms, with support for 129 languages. It is tuned for the model the user actually uses; each prompt receives a 16-dimension Prompt Score, and skills receive a Skill Score. Saved prompts are stored in PromptDrive and can be used across platforms. The service is free to start.

Main capabilities:

• Generate contextual prompts and agent skills for more than 140 AI platforms, covering 129 languages.
• Each prompt provides a 16-dimension Prompt Score; skills receive a Skill Score in workflows (score_skill is not a separately exposed MCP tool).
• Saved prompts are stored in PromptDrive and can be used across platforms.
• Generation and scoring may consume account quota; successfully generated results are automatically saved to PromptDrive, and the actually returned savedToVault and promptId fields indicate the persisted result.
• The Memory tool can access the user's Prompeteer Memory when enabled and authorized (referring to the source stored by Prompeteer, not the host assistant's memory or conversation history).

How to use:

1. Connect to the hosted server: use a client that supports remote MCP and OAuth discovery, add the endpoint in the remote server settings, and complete browser login. Access depends on the account, plan, consented scopes, and client support.
2. Connect via the npm stdio bridge: suitable for clients that launch local MCP commands, and requires Node.js 22.14.0 or later. This command runs a stdio bridge to the same hosted Prompeteer endpoint, using a pinned version of the mcp-remote 0.8.6 proxy to handle Streamable HTTP and OAuth. The client opens browser login when authorization is required. The local bridge does not run Prompeteer's generation models or product services.
3. The endpoint and OAuth resource are pinned to https://prompeteer.ai/mcp。包装器使用回环 OAuth callback, no shell execution, and forwards MCP messages via stdout; connection diagnostics use stderr. It does not accept arbitrary endpoints, authentication headers, or debug flags. Native remote clients can use the URL directly without installing the bridge.
4. --json prints Cursor's direct remote configuration; --help and --version do not start a connection.
5. OAuth credentials are managed by mcp-remote, using the local ~/.mcp-auth cache by default. On POSIX systems, new credential files are written with mode 0600. The bridge does not add its own token storage.
6. Example: "Generate a prompt for a product launch email in Claude. Show its Prompt Score, then save the approved version to PromptDrive."

Repository manifest and plugins:

• plugin.json and mcp.json provide the portable Agent Plugins format supported by Cursor. After installing the plugin through Cursor, connect to Prompeteer and complete browser login.
• gemini-extension.json provides a Gemini CLI extension using the hosted HTTP endpoint and OAuth discovery.
• .claude-plugin/plugin.json provides plugin metadata, and .mcp.json connects clients directly via HTTP to https://prompeteer.ai/mcp。该插件包含一个远程 MCP connection, without adding local commands, hooks, skills, agents, or shell execution, and without starting the npm bridge or installing its runtime dependencies.
• These manifests have passed structural validation, but installation and OAuth in the actual Cursor and Gemini CLI applications have not yet been tested for this version.

Network access and account permissions: The client will contact https://prompeteer.ai/mcp as well as Prompeteer at https://prompeteer.ai OAuth discovery, registration, authorization, and token endpoints on the site. Prompeteer sign-in may redirect to the identity provider chosen by the user. Credentials remain within the OAuth flow supported by the client; the plugin does not contain tokens, API keys, or environment variable readers. A Prompeteer account is required. Tool availability depends on the authenticated account, approved scopes, and plan quota.

Build and verification: Use Node.js 22.14.0 or later. The source code locks the maintained proxy and its dependencies. The build copies unmodified runtime files and their licenses into the release package. The explicit build uses mcp-remote 0.8.6 and patched qs 6.16.0. Tests run protocol forwarding and startup disconnect checks against local fixtures, without generating prompts or making paid service calls. The vendor/ build output is included in the npm release and is not committed to this repository.

Registration and origin: The official MCP registry lists the server as io.github.prompeteer/prompeteer-mcp. server.json describes the npm stdio bridge and the hosted Streamable HTTP endpoint. Version 2.0.3 adds a working stdio bridge and corrects outdated setup instructions.

Support and licensing: The bridge wrapper and the documentation in this repository are licensed under MIT. The hosted Prompeteer platform is proprietary. Contact info@prompeteer.com for support or privacy questions. Disconnect the MCP connection in the client when it is no longer needed.

Pricing and public facts: Start for free. Prompeteer reports having 12,000+ users in 160 countries and supporting 129 languages. These company facts were approved for release in September 2026.