spikeme-mcp

spikeme-mcp

The SpikeMe MCP server. It speaks the Model Context Protocol, so a coding agent like Claude Code or Cursor can call SpikeMe directly — analyze your repo's stack and generate a grounded, time-boxed spike decision document without leaving the editor.

Tools

• analyze_stack — analyze the repo manifest (package.json, pyproject.toml, requirements.txt) and return the stack, categories, and the gaps that make good spike topics. Local and free, no account.
• generate_spike — generate a full spike decision document for a topic, grounded in the repo's real stack and in live registry facts (npm/PyPI). Requires a SpikeMe login and an active plan.

Quick start

Claude Code

claude mcp add spikeme -- npx -y spikeme-mcp

Cursor and other MCP clients

Add it to your MCP config (e.g. .cursor/mcp.json or your client's mcpServers block):

{
  "mcpServers": {
    "spikeme": {
      "command": "npx",
      "args": ["-y", "spikeme-mcp"]
    }
  }
}

Authentication

generate_spike needs an active plan. Sign in once with the SpikeMe CLI — npx spikeme login — and the server reuses those credentials (~/.spikeme/credentials.json). For headless setups, set the SPIKEME_TOKEN environment variable in the server's env instead. analyze_stack is always free and needs no token.

Docs

Full documentation: https://spikeme.io/docs/mcp

The CLI (same package family) lives at spikeme.

MIT © SpikeMe