Brainstorm MCP
VerifiedA single model's confident answer can still be wrong, so brainstorm-mcp runs multi-round debates between GPT, Gemini, DeepSeek, Claude, and local Ollama models and returns a structured synthesis, as an MCP server that needs no API keys in hosted mode.
Set up this server
01 / Choose your client
02 / Before you connect
Authentication is not specified. Check the project instructions before connecting.
Project instructions03 / Add the configuration
claude mcp add --transport stdio 'brainstorm-mcp' -- 'npx' '-y' 'brainstorm-mcp'04 / Check it in your client
Open your client’s MCP settings and confirm the server connects and lists its tools. A copied configuration does not confirm a working connection.
More in AI & ML
Browse the full directoryBadge
Show this listing on your README and website
You are one of MCPVault's first 100 verified servers: verified for good, badge or no badge. Put it on your README or site if you want a live Verified proof link for visitors.
Security profile
Exposure: Limited
Runs commands or loads code on your machine, but does not reach secrets or other agents.
What this server can reach, from a static read of the shipped code. Nothing was executed and nothing here is a verdict: a browser tool has to run commands and an API wrapper has to call its API. Decide what fits your setup.
Detected
- Talks to external services
Makes outbound requests. The hosts are listed so you can see where data goes.
api.groq.com, api.minimax.io, api.mistral.ai, api.moonshot.ai, api.openai.com, api.together.xyz, api.z.ai
- Contains instructions aimed at AI agents
Text in code or agent-facing files that steers a model. Worth reading before you trust it.
Dependencies with known advisories
- @hono/node-server 1.19.9 High, CVE-2026-39406, GHSA-FRVP-7C67-39W9, CVE-2026-29087
- express-rate-limit 8.2.1 High, CVE-2026-30827
- fast-uri 3.1.0 High, CVE-2026-13676, CVE-2026-18446, CVE-2026-75975
- hono 4.12.1 High, CVE-2026-56762, CVE-2026-47676, CVE-2026-47675
- ip-address 10.0.1 High, CVE-2026-69192, CVE-2026-42338
- path-to-regexp 8.3.0 High, CVE-2026-4923, CVE-2026-4926
- body-parser 2.2.2 Low, CVE-2026-12590
- esbuild 0.27.3 Low, GHSA-G7R4-M6W7-QQQR
Left out of the profile
24 pattern matches in documentation, configuration or CI files were recorded but are not part of the profile, because they do not run when the server does. The owner dashboard lists them.
7 of 7 tools · Observed tool names: 7/7
Observed tool names (7)
Reported by a complete handshake for the current saved configuration. Tools were listed, not invoked.
· 1.3.0
add_providerbrainstormbrainstorm_collectbrainstorm_quickbrainstorm_respondbrainstorm_reviewlist_providersDocumented tools (7)
From project documentation. A server handshake does not verify each tool’s description or behavior.
add_provider
Add a new API or CLI provider at runtime
brainstorm
Multi-round debate between AI models (API or hosted mode)
brainstorm_collect
Submit model responses in a hosted session
brainstorm_quick
Instant multi-model perspectives — parallel, no rounds
brainstorm_respond
Submit Claude's response in an interactive session
brainstorm_review
Multi-model code review with findings, severity, verdict
list_providers
Show configured providers, API key status, and detected CLIs
Tool change history
FAQ
Questions about Brainstorm MCP Server
- How do I connect Brainstorm MCP Server to Claude?
- Run `claude mcp add brainstorm-mcp -- npx -y brainstorm-mcp` in Claude Code, or add the same command and arguments under mcpServers in Cursor's mcp.json or Claude Desktop's claude_desktop_config.json, then restart the client. The blocks above are ready to paste.
- Is Brainstorm MCP Server free?
- The listed licence is MIT. Check the upstream terms for permitted use and commercial requirements; a public repository does not by itself mean the software is free or open source. Connected APIs and hosted services may have separate charges.
- What can Brainstorm MCP Server do?
- Brainstorm MCP Server documents 7 tools to the agent, including add_provider, brainstorm, brainstorm_collect. The descriptions above come from project documentation. A live handshake does not test individual tool behavior.