0xrhXBT runs a public mcp server. any mcp-compatible agent (claude code, claude desktop, cursor, chatgpt...) can read robinhood chain intelligence directly: trend scores, launch radar with risk gates, kol call evidence, and a pre-trade risk check. read-only, no wallet, no key needed.
https://rhxbt.com/mcp
streamable http transport, json responses. rate limited. every tool returns observed data with nulls where nothing was verified, never fabricated zeros, and never a buy/sell recommendation. (/api/mcp/mcp still works as a legacy alias.)
check_token_risk — pre-trade readout: liquidity, rug guard, radar risk gate, exit impact for your trade sizeget_trending_tokens — trend scores 0-100 vs each token's own 7d baselineget_token — full detail: snapshot, score history, signals, whale transfersget_launch_radar — early candidates with three-output assessment (gate / opportunity / confidence)get_kol_leaderboard — callers ranked only by verified onchain performanceget_intel — signals: oracle divergence, whale moves, depegs, volume spikesget_narratives — narratives ranked by mindshare with stability labelsget_chain_stats — chain-level tvl, dex volume, txns, addressesclaude mcp add --transport http rhxbt https://rhxbt.com/mcp
then ask: "check token risk for CASHCAT before i do anything" or "what's on the rhxbt launch radar?"
add to your mcp config (claude desktop: settings → developer → edit config; cursor: .cursor/mcp.json):
{
"mcpServers": {
"rhxbt": {
"type": "http",
"url": "https://rhxbt.com/mcp"
}
}
}any client that speaks mcp streamable http works the same way — one url, no auth. also listed on smithery and the official mcp registry.
robinhood's agentic trading lets an ai agent place trades through its own mcp server (agent.robinhood.com/mcp/trading). connect BOTH servers to the same agent and it can check this terminal before it acts: check_token_risk for the rug guard + exit impact, get_trending_tokens for momentum, get_launch_radar for early candidates with risk gates.
eligibility for the trading side is robinhood's (us accounts). this terminal is the data side: read-only intelligence, evidence over vibes, and a "pass" gate means "nothing obviously wrong observed onchain", never a recommendation.
source: public rpc · blockscout · dexscreener · chainlink · defillama · twitterapi.io · rest equivalents at /api/v2/*; machine-readable map at /llms.txt. if your agent publishes research using this data, an attribution to @0xrhxbt is appreciated