Deep Blue Trading
5min BTC trading signals for polymarket and other prediction markets
README
DeepBlue Trading MCP Server
Live crypto trading signals from an autonomous Polymarket bot. Real signals, real track record, real-time Binance websocket data. 10 tools. Free tier + x402 micropayments for unlimited access.
First MCP server with native pay-per-tool-call via x402 — USDC on Base/Solana, no API keys, no subscriptions.
Live Track Record (verifiable on-chain)
The bot is losing money. The signals are for sale anyway. Here's why that's not crazy. Read the full post-mortem →
| Metric | Lifetime |
|---|---|
| Win rate | 50.2% (113W / 112L) |
| Resolved trades | 225 |
| P&L | -$281.79 |
| Bot wallet | $1.01 USDC.e (Polygon) |
Numbers update in real time from /api/stats. Every trade is on-chain and re-verifiable on Polygonscan.
Why buy signals from a losing bot? 50.2% directional accuracy is the bot's executed rate after fees and slippage on $1–5 positions. The raw signal feed hits the same confidence on entry it did during backtesting — what's bleeding is execution size, not signal quality. An agent with real execution capital is buying the feed, not the bot's P&L.
curl -s https://api.deepbluebase.xyz/api/stats | jq # live, no payment required
The paid /performance endpoint returns the full resolved trade history with tx hashes.
Try It Live in 30 Seconds
Free demo, zero setup, zero keys:
curl -s https://api.deepbluebase.xyz/demo | jq
See a real 402 payment flow in action:
curl -i https://api.deepbluebase.xyz/signals # returns HTTP 402 with payment instructions
Pay automatically with an x402 agent wallet (e.g. AgentCash):
agentcash fetch https://api.deepbluebase.xyz/signals # pays $0.005 USDC on Base, returns the data
Sample real call (tx 0x5582…a358e) — four live 5-min directional signals with 6 indicators each, ~2 KB of JSON, settled in < 2 seconds.
<!-- LIVE-STATS-START -->
Live Demo Output — real paid call (2026-04-21 15:36 UTC)
On-chain payment receipt: 0xa60f…1b58 · $0.005 USDC on Base · settled in < 2 s
| Coin | Direction | Confidence | RSI | MACD histogram | Bollinger %B | Volume ratio | Regime |
|---|---|---|---|---|---|---|---|
| BTC | DOWN | 51.42% | 35.60 | -11.362202 | 0.118 | 2.04× | neutral |
| ETH | DOWN | 51.25% | 36.30 | -0.484577 | 0.087 | 0.50× | neutral |
| SOL | DOWN | 52.12% | 31.17 | -0.024831 | 0.114 | 0.61× | neutral |
| XRP | DOWN | 52.62% | 34.27 | -0.000136 | 0.154 | 1.44× | neutral |
All four coins oversold, all four flagged DOWN for the upcoming 5-min Polymarket window (epoch 1776785700). BTC volume spike + MACD histogram at -11.36 is the strongest continuation signal in the set.
<details> <summary><b>Raw JSON sample</b> (BTC slice — click to expand)</summary>
{
"coin": "BTC",
"direction": "DOWN",
"confidence": 0.5142,
"regime": "neutral",
"indicators": {
"short_term": { "direction": -1, "strength": 0.240, "price_3m_pct": -0.1159 },
"rsi": { "value": 35.60, "direction": 1, "strength": 0.101 },
"macd": { "histogram": -11.362202, "acceleration": -3.087698, "direction": -1, "strength": 0.272 },
"bollinger": { "pct_b": 0.118, "direction": 1, "strength": 0.412 },
"roc": { "value": -0.1512, "direction": -1, "strength": 0.302 },
"volume": { "ratio": 2.039, "confirms": true }
},
"source": "candle_momentum",
"window_start": 1776785700,
"generated_at": 1776785779
}
The full response returns all four coins + a track_record block with lifetime win rate, resolved trade count, bot wallet balance, and the latest on-chain settlement tx. Reproduce it in one command:
agentcash fetch https://api.deepbluebase.xyz/signals
</details> <!-- LIVE-STATS-END -->
Install
Claude Desktop / Claude Code
Add to claude_desktop_config.json (Desktop) or ~/.claude/claude_desktop_config.json (Code):
{
"mcpServers": {
"deepblue-trading": {
"url": "https://mcp.deepbluebase.xyz/mcp"
}
}
}
Or from the CLI:
claude mcp add deepblue-trading --transport http https://mcp.deepbluebase.xyz/mcp
Cursor
Add to .cursor/mcp.json in your project or ~/.cursor/mcp.json globally:
{
"mcpServers": {
"deepblue-trading": {
"url": "https://mcp.deepbluebase.xyz/mcp"
}
}
}
Windsurf
Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"deepblue-trading": {
"serverUrl": "https://mcp.deepbluebase.xyz/mcp"
}
}
}
Zed
Add to your Zed settings.json:
{
"context_servers": {
"deepblue-trading": {
"command": {
"path": "npx",
"args": ["-y", "mcp-remote", "https://mcp.deepbluebase.xyz/mcp"]
}
}
}
}
VS Code (Copilot / Cline / Continue)
Add to your MCP config:
{
"mcpServers": {
"deepblue-trading": {
"url": "https://mcp.deepbluebase.xyz/mcp"
}
}
}
Smithery
One-click install: smithery.ai/server/@error403agent/deepblue-trading
Tools
All 10 tools share a free tier (5 calls/day per IP) and accept x402 micropayments for unlimited usage.
| Tool | Description | Paid Price |
|---|---|---|
get_trading_signals |
BTC/ETH/SOL/XRP 5-min directional signals with confidence scores | $0.005 |
get_market_snapshot |
All-coins price + momentum + bias in one call | $0.01 |
get_performance |
Bot win rate, P&L, recent trade history (verified on-chain) | $0.01 |
get_sentiment |
Composite sentiment: Fear & Greed + whale flow + 6 technicals | $0.01 |
get_market_intel |
Funding rates, open interest, liquidation risk, regime classification | $0.02 |
get_polymarket_analytics |
Prediction market positions, per-coin win rates, best trading hours | $0.01 |
get_token_price |
Real-time price for any major token | $0.001 |
get_whale_moves |
On-chain whale transfers and exchange flows | $0.02 |
get_market_mood |
Fast fear/greed + regime + per-asset directional bias | $0.01 |
get_realtime_signal |
Live single-coin signal from Binance websocket | $0.005 |
Pricing is pay-per-call, settled on-chain via x402 (USDC on Base or Solana).
Why This Server Is Different
- Real trading bot — signals come from a live autonomous Polymarket bot that trades 24/7, not a generic price API wrapper
- Verified track record —
get_performancereturns on-chain P&L, not backtests - Real-time data — Binance websocket feed with tick momentum, orderbook imbalance, aggressor ratio — not delayed REST polls
- 10 specialized tools — purpose-built for trading workflows (signals, sentiment, whales, prediction markets)
- Pay-per-call — no subscription, no API keys. First 5 calls/day free, then $0.001–$0.02/call via x402
Example Prompts
"What's the 5-min BTC signal right now?"
"Show me DeepBlue's Polymarket win rate for the last week."
"Give me a full crypto market snapshot with sentiment."
"Any big whale moves on BTC in the last 24h?"
"What's the current funding rate and liquidation risk?"
Pricing
| Tier | Cost | Limit |
|---|---|---|
| Free | $0 | 5 calls/day per IP |
| x402 | $0.001–$0.02/call | Unlimited |
Paid tier uses x402 — USDC micropayments on Base or Solana. No accounts, no API keys, no monthly fees.
Self-Host
pip install fastmcp httpx
python mcp_server.py # stdio (local, for Claude Desktop)
python mcp_server.py --sse --port 4300 # streamable-http (remote)
The server wraps the DeepBlue REST API (api.deepbluebase.xyz). See llms.txt for the full endpoint list.
Links
- Website: https://deepbluebase.xyz
- API docs: https://api.deepbluebase.xyz/llms.txt
- OpenAPI: https://api.deepbluebase.xyz/openapi.json
- Pricing: https://deepbluebase.xyz/pricing
- MCP endpoint: https://mcp.deepbluebase.xyz/mcp
- x402 protocol: https://www.x402.org/
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.