idea-reality
Pre-build reality check for AI coding agents. Scans GitHub, HN, npm, PyPI & Product Hunt — returns a 0-100 reality signal. MCP tool. Try: mnemox.ai/check
claude mcp add --transport stdio mnemox-ai-idea-reality-mcp uvx idea-reality-mcp \ --env GITHUB_TOKEN="your_github_token_here (optional - increases rate limits)" \ --env PRODUCTHUNT_TOKEN="your_producthunt_token_here (optional - enable deep mode)"
How to use
idea-reality-mcp is a Python-based MCP server that performs a reality check for a given idea by querying five real data sources in parallel: GitHub repos, Hacker News discussions, npm packages, PyPI packages, and Product Hunt. It returns a reality_signal score along with supporting evidence, top similars, and pivot hints to help your agent decide whether to build, pivot, or partner. To use it, run the MCP server via your MCP client (the recommended invocation is uvx idea-reality-mcp). You can then prompt your agent to call idea_check with a natural-language description of the idea. The tool will automatically fetch sources and provide structured outputs that your agent can interpret for decision making. You can also integrate the tool into CI or other tooling via your MCP client of choice (Claude, Cursor, Code, etc.).
How to install
Prerequisites:
- Python 3.11 or newer
- Optional: pipx for isolated install
- Optional: pip if you prefer direct install
Install methods:
- Using pip (system-wide or virtualenv):
python -m ensurepip --upgrade
pip install --upgrade idea-reality-mcp
- Using pipx (isolated, recommended):
python -m pip install --user pipx
python -m pipx ensurepath
pipx install idea-reality-mcp
- Verify installation:
python -m idea_reality_mcp --version # if the package provides a CLI entry
# or depending on package, simply run the MCP via your MCP client as documented
Usage after installation:
- Run the MCP server via your MCP client, e.g., uvx idea-reality-mcp (as shown in the README).
- Ensure your environment variables (e.g., GITHUB_TOKEN) are set if you plan to access authenticated sources.
Additional notes
Notes and tips:
- The server pulls from five sources in parallel to produce a reality_check result with a reality_signal (0-100) and supporting evidence.
- If you hit API rate limits, provide GITHUB_TOKEN (and any other tokens you use) to increase limits.
- This MCP is open-source and free to use; you can audit and customize the sources or scoring if needed.
- When configuring in Claude Desktop or other clients, ensure the mcpServers entry matches your client’s expected format.
- For CI or automation, consider proactive triggering by adding a line to your CLAUDE.md or equivalent to call idea_check before starting a project.
Related MCP Servers
gemini -tool
MCP server that enables AI assistants to interact with Google Gemini CLI, leveraging Gemini's massive token window for large file analysis and codebase understanding
python-utcp
Official python implementation of UTCP. UTCP is an open standard that lets AI agents call any API directly, without extra middleware.
neurolink
Universal AI Development Platform with MCP server integration, multi-provider support, and professional CLI. Build, test, and deploy AI applications with multiple ai providers.
go-utcp
Official Go implementation of the UTCP
crash
MCP server for structured and efficient reasoning with step validation, branching, and revisions.
rs-utcp
Official Rust implementation of the UTCP