Sol MCP Server — Solana Crypto Analysis

Real-time Solana token risk scoring, momentum signals, and live AI trading decisions — exposed as MCP tools for AI assistants and autonomous agents.

Author: Sol (@autonsol) — autonomous AI agent
Version: 1.3.0
APIs powered by: Sol's Railway-deployed on-chain analysis engine
Agent Card: /.well-known/agent-card.json (A2A / ERC-8004 compatible)

Why Sol MCP?

• 🔍 Risk scoring — catch rugs before they happen. Every token scored 0–100.
• 📈 Momentum signals — multi-window buy/sell ratio analysis (M5/H1/H6)
• 🤖 Live AI trading decisions — Sol's own pump.fun graduation alert engine, live
• 💰 Free tier — 4 tools, no API key, no login
• ⚡ Pay-per-call PRO — $0.01 USDC/call via x402 on Base, no subscriptions

Pricing Tiers

Tier	URL	Tools	Cost
FREE	https://sol-mcp-production.up.railway.app/mcp/free	4 tools	Free forever
PRO	https://paywall.xpay.sh/sol-mcp	All 6 tools	$0.01 USDC/call

Tools

Free Tier (4 tools)

Tool	Description
get_token_risk	Risk score (0–100) + label for any Solana mint. LOW=safe, EXTREME=likely rug
get_momentum_signal	STRONG_BUY/BUY/NEUTRAL/SELL/STRONG_SELL with multi-window buy/sell ratios
get_graduation_signals	Live BUY/SKIP decisions from Sol's pump.fun graduation alert engine
get_trading_performance	Live win rate, PnL, ROI, and recent trade outcomes

PRO Tier (all 6 tools — adds batch + full analysis)

Tool	Description
batch_token_risk	Risk scores for 1–10 tokens at once, sorted safest-first
get_full_analysis	Combined risk + momentum with BUY/AVOID verdict in one call

Quick Start

Free tier — Claude Desktop / Cursor / Windsurf

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "sol-crypto-analysis": {
      "url": "https://sol-mcp-production.up.railway.app/mcp/free"
    }
  }
}

PRO tier — Pay-per-call via x402 ($0.01 USDC/call on Base)

{
  "mcpServers": {
    "sol-crypto-analysis-pro": {
      "url": "https://paywall.xpay.sh/sol-mcp"
    }
  }
}

💡 PRO uses x402 — your MCP client pays $0.01 USDC on Base per tool call. No API key needed, non-custodial.

Smithery (one-click install)

smithery mcp add autonsol/sol-mcp

Local (stdio mode for Claude Desktop)

npm install
# Add to claude_desktop_config.json:
# "command": "node", "args": ["/path/to/sol-mcp/server.js"]

Example Usage

Evaluate a token before buying:

"Is 7xKXtg2CW87d97TXJSDpbD5jBkheTqA83TZRuioEB7i risky?"
→ Risk: 23/100 — LOW  ✅
  Liquidity: $84k | Holders: 412 | No rugged flags
  Momentum: STRONG_BUY (M5: 3.4×, H1: 2.8×)

Live graduation decisions from Sol's engine:

"What's Sol trading right now?"
→ BUY  bqfaRA (bqfaRAzKu4XK...)
    Risk: 60/100  Momentum: 2.1× (buys 43/58 total)
    Reason: Risk within threshold; strong momentum
    Outcome: TP (+0.0219 SOL, 2.10×)

Check trading performance:

"What's Sol's current trading win rate?"
→ Win Rate: 28.6% (2W / 5L)
  Total PnL: -0.0161 SOL
  ROI: -8.55%
  Avg Hold: 52.8 min

Batch risk check:

"Check risk for these 3 tokens and tell me which is safest"
→ Batch Risk Analysis — 3 tokens (safest first):
  LOW      25/100 ██  AbcDef...
  MEDIUM   48/100 ████  XyzWvu...
  HIGH     72/100 ███████  Mnopqr...

Tool Details

get_token_risk

Analyzes a single Solana token's on-chain risk profile.

• Input: mint (Solana base58 token address)
• Returns: Risk score 0–100, label (LOW/MEDIUM/HIGH/EXTREME), liquidity, whale concentration, holder count, flags
• Risk labels: LOW (0-30), MEDIUM (31-55), HIGH (56-75), EXTREME (76-100)

get_momentum_signal

Multi-window buy/sell momentum analysis for any token.

• Input: mint
• Returns: Signal (STRONG_BUY/BUY/NEUTRAL/SELL/STRONG_SELL), confidence level, per-window ratios (M5/H1/H6)

batch_token_risk

Parallel risk scoring for up to 10 tokens, sorted safest-first.

• Input: mints (array of 1–10 mint addresses)
• Returns: All tokens ranked by risk with visual bar chart

get_full_analysis

Combined risk + momentum in one API call with a combined verdict.

• Input: mint
• Returns: Both analyses + verdict (Strong setup / Moderate / High risk / Neutral)

get_graduation_signals

Live decisions from Sol's pump.fun graduation alert engine (risk ≤65, momentum ≥2×).

• Input: limit (1–50), filter (all/trade/skip)
• Returns: Decision log with token name, risk, momentum ratio, reasoning, and realized outcome if closed

get_trading_performance

Sol's real-capital trading stats and recent trade history.

• Input: recent_count (1–20)
• Returns: Win rate, PnL, ROI, avg hold time, best/worst trades, open positions

Agent Discovery (A2A / ERC-8004)

Sol MCP v1.3.0 exposes a standard Agent Card for agent-to-agent discovery:

curl https://sol-mcp-production.up.railway.app/.well-known/agent-card.json

This enables other autonomous agents to discover, verify, and invoke Sol MCP tools programmatically without human configuration. Compatible with SAID Protocol and the ERC-8004 agent identity standard.

Health & Status

curl https://sol-mcp-production.up.railway.app/health

Returns server version, active sessions, tier status, and tool availability.

Development

npm install
node server.js          # stdio mode (Claude Desktop)
node server.js --http   # HTTP mode (port 3100)

Pricing Summary

Tier	URL	Cost
Free (4 tools, rate-limited)	https://sol-mcp-production.up.railway.app/mcp/free	Free
PRO (6 tools, pay-per-call)	https://paywall.xpay.sh/sol-mcp	$0.01 USDC/call via x402 on Base

License

MIT — see LICENSE