CrewAI MCP Crypto Intel Agents

A multi-agent system that generates institutional-grade cryptocurrency market intelligence reports with real-time price data, on-chain analytics, macroeconomic context, and social sentiment — powered by CrewAI, Google Gemini, and 15 production MCP servers from the Vinkius AI Gateway.

---

Why Crypto Intelligence Needs Real-Time Data Sources

The cryptocurrency market operates 24/7 across thousands of trading pairs, dozens of blockchains, and millions of social media conversations. Making informed decisions requires synthesizing data from fundamentally different domains — exchange prices, blockchain network metrics, Federal Reserve monetary policy, SEC filings, and Crypto Twitter sentiment — often within a window of hours.

Most crypto analysis available online falls into two categories: surface-level price commentary that adds nothing, or paywalled institutional research that costs $50,000+ per year. There is almost nothing in between — no publicly available analysis that combines verified exchange data with on-chain metrics, macro correlations, and quantified social sentiment in a single coherent report.

This project fills that gap. Four specialized AI agents collaborate to produce an institutional-quality intelligence report for any cryptocurrency, using real-time data from 15 MCP servers. The agents do not speculate or fabricate — they query live market data, blockchain explorers, economic databases, and social platforms, then synthesize their findings into a scenario-based analysis with proper risk disclaimers.

---

How It Works

The system operates as a sequential pipeline of four agents, each connected to curated MCP servers hosted on the Vinkius AI Gateway.

Phase 1 — Market Data Collection

The first agent builds a quantitative market profile. It connects to five MCP servers:

• CoinGecko for comprehensive market data, historical prices, and global metrics
• CoinMarketCap for rankings, market pairs, and exchange volume distribution
• Binance for real-time order book data, recent trades, and ticker information
• Alpha Vantage for technical indicators, forex rates, and equity correlations
• Twelve Data for time series analysis and cross-asset technical comparisons

The output includes current prices cross-referenced across exchanges, RSI, MACD, moving averages, support/resistance levels, and volatility metrics — all with source attribution.

Phase 2 — On-Chain and Macroeconomic Analysis

The second agent connects crypto to the real world. It uses four MCP servers:

• Blockchain.com for on-chain metrics — active addresses, transaction counts, hash rate, whale movements, exchange flows
• SEC EDGAR for regulatory filings — ETF applications, enforcement actions, insider trades at crypto companies
• FRED (Federal Reserve Economic Data) for macro indicators — Fed funds rate, CPI, DXY, yield curves
• CoinMarketCal for the crypto event calendar — token unlocks, protocol upgrades, halvings, conference dates

This phase answers the question institutional investors always ask: "What is the macro context, and is the regulatory environment getting better or worse?"

Phase 3 — Sentiment Intelligence

The third agent mines social media and news. It connects to four MCP servers:

• X/Twitter for Crypto Twitter mentions, volume trends, and key influencer positions
• NewsAPI for crypto news headlines from the past 7 days, categorized by sentiment
• Exa AI for semantic search across Reddit, Telegram summaries, and crypto forums
• CoinDesk for the Bitcoin Price Index and market reference data

The output quantifies sentiment: "Social sentiment shifted from 65% bullish to 48% over the past 72 hours, driven primarily by regulatory uncertainty in 3 jurisdictions."

Phase 4 — Intelligence Report

The fourth agent compiles everything into a structured report. It uses two MCP servers for SEO optimization:

• SEMrush for crypto keyword research and content gap analysis
• SerpAPI for current SERP landscape and competitor report analysis

The final report includes scenario analysis (bull/base/bear cases with probability weightings), ranked risk factors, key levels to watch, and an upcoming catalyst calendar — all framed with appropriate disclaimers, never as predictions.

---

Why MCP Servers Are Essential for Financial Intelligence

Building a crypto intelligence system without MCP would require custom integrations with 15 different APIs — each with authentication, rate limiting, response parsing, and maintenance. The CoinGecko API alone has had 4 breaking changes in the past year. Maintaining all of these simultaneously is a full-time engineering job.

The Vinkius AI Gateway provides over 2,600 production-ready MCP servers that handle all of this complexity. Each server is an authenticated SSE endpoint:

https://edge.vinkius.com/<your_token>/mcp

This project uses 15 of those 2,600+ servers. The financial data coverage on the Vinkius AI Gateway extends far beyond what this project uses:

• Additional crypto data via Coinbase, EODHD Financial, and Flux Markets
• Traditional finance via Finnhub, SEC EDGAR Financials, and ECB Eurozone Intelligence
• Economic indicators via FRED Series, FRED GeoFRED, and US Treasury
• Commodity markets via EIA Petroleum and EIA Natural Gas

Adding any of these is a one-line configuration change. Browse the full catalog at vinkius.com/en/categories.

---

MCP Servers Used in This Project

| MCP Server | Agent | Data Provided | |---|---|---| | coingecko-mcp | Market Data | Prices, market caps, volumes, historical data | | coinmarketcap-mcp | Market Data | Rankings, market pairs, exchange volumes | | binance-crypto-market-mcp | Market Data | Real-time order books, trades, tickers | | alpha-vantage-mcp | Market Data | Technical indicators, forex, equities | | twelve-data-mcp | Market Data | Time series, cross-asset technicals | | blockchaincom-data-mcp | On-Chain & Macro | Block explorer, wallet data, network stats | | sec-edgar-full-the-ultimate-free-bloomberg-alternative-for-ai-agents-mcp | On-Chain & Macro | SEC filings, insider trades, enforcement | | fred-full-access-us-economic-intelligence-mcp | On-Chain & Macro | Fed rate, CPI, GDP, yield curves | | coinmarketcal-mcp | On-Chain & Macro | Crypto event calendar, token unlocks | | x-twitter-mcp | Sentiment | Crypto Twitter, influencer positions | | newsapi-mcp | Sentiment | Crypto news headlines, outlet coverage | | exa-ai-mcp | Sentiment | Reddit, forums, semantic search | | coindesk-bitcoin-price-index-mcp | Sentiment | CoinDesk BPI, market reference data | | semrush-mcp | Report Writer | Crypto keyword research | | serpapi-mcp | Report Writer | SERP analysis, competitor reports |

All 15 servers are hosted on the Vinkius AI Gateway. Browse the full catalog of 2,600+ production-ready MCP servers at vinkius.com/en/categories.

---

Getting Started

Prerequisites

• Python 3.11+
• A Google Gemini API key (free tier is sufficient)
• MCP server endpoints from the Vinkius AI Gateway

Installation

git clone https://github.com/vinkius-labs/crewai-mcp-crypto-intel-agents.git
cd crewai-mcp-crypto-intel-agents

python -m venv .venv
source .venv/bin/activate    # Linux/macOS
# .venv\Scripts\activate     # Windows

pip install -e .

Configuration

cp .env.example .env

Open .env and add your Gemini API key and Vinkius MCP URLs from the Vinkius AI Gateway marketplace.

Usage

# Validate your environment
crypto-intel validate

# Generate a Bitcoin intelligence report
crypto-intel analyze "Bitcoin" "BTC"

# Generate an Ethereum report
crypto-intel analyze "Ethereum" "ETH"

# Generate a Solana report
crypto-intel analyze "Solana" "SOL"

---

Generated Report Structure

| Section | Content | |---|---| | Executive Summary | 3-sentence asset overview, key signal, primary risk | | Price Action & Technicals | Current prices, RSI, MACD, MAs, support/resistance | | On-Chain Health | Active addresses, whale flows, exchange net flows | | Macro Environment | Fed rate, CPI, DXY correlation, equity comparison | | Market Sentiment | Quantified social sentiment, news analysis, community quotes | | Regulatory Landscape | SEC filings, enforcement actions, legislative signals | | Scenario Analysis | Bull / Base / Bear cases with probability weightings | | Risk Factors | Top 5 risks ranked by probability and impact | | Key Levels to Watch | Support, resistance, and trigger points | | Upcoming Catalysts | Events calendar with potential impact ratings | | Disclaimer | Standard financial research disclaimer |

---

Technical Details

• Framework: CrewAI with Flows and @CrewBase decorators
• LLM: Google Gemini 2.0 Flash (free tier, ~15 RPM)
• State Management: Pydantic models for type-safe data flow between agents
• MCP Integration: Native CrewAI mcps= field with SSE transport to Vinkius AI Gateway
• CLI: Typer with Rich console output
• Rate Limiting: max_rpm=10 per agent to stay within Gemini free tier limits

---

Disclaimer

This tool generates research reports for informational purposes only. It does not constitute financial advice, investment recommendations, or trading signals. Always conduct your own research and consult with qualified financial advisors before making investment decisions. Cryptocurrency investments carry significant risk, including the potential loss of all invested capital.

---

FAQ

What is MCP?

The Model Context Protocol is an open standard for connecting AI systems to external tools and data sources. See modelcontextprotocol.io.

Can I add more financial data sources?

Yes. The Vinkius AI Gateway offers MCP servers for Coinbase, Finnhub, SEC EDGAR Financials, ECB Eurozone, and hundreds more — explore the full catalog.

Can I use a different LLM?

Yes. CrewAI supports OpenAI, Anthropic, Mistral, and any LiteLLM-compatible model.

Does this provide trading signals?

No. This tool generates research reports. It never uses "guaranteed" language and always presents scenario-based analysis with probability weightings and risk disclaimers.

---

Contributing

We welcome contributions. Please read the Contributing Guide before submitting a pull request.

---

License

MIT — see LICENSE.

---

Built by Vinkius Labs with CrewAI and the Vinkius AI Gateway.