Scout Intel MCP — The Google for AI Agents

Scout MCP
Business & Market Intelligence for AI Agents

Google for AI agents— instead of web pages, it returns clean, structured JSON that agents can reason over.

Scout MCP gives any AI agent instant access to structured business intelligence, market research, and competitive analysis. It aggregates data fromDuckDuckGo, NewsAPI, Wikipedia, web scraping, and social profilesinto Pydantic-validated JSON responses withper-source confidence breakdownsanddata quality grades.

- Quick Install
- The 6 Intelligence Tools
- Data Quality Grades
- Confidence Breakdown
- Full API Reference
- Example Responses
- Architecture
- Configuration
- Pricing & Rate Limits
- Self-Hosting
- Docker
- Tech Stack
- Data Sources
- Contributing

{ "mcpServers": { "scout-mcp": { "command": "python", "args": ["-m", "scout_mcp.mcp_server"], "cwd": "/path/to/scout-mcp/src", "env": { "NEWS_API_KEY": "your-newsapi-key" } } } }

{ "scout-mcp": { "command": "python", "args": ["-m", "scout_mcp.mcp_server"], "cwd": "/path/to/scout-mcp/src", "env": { "NEWS_API_KEY": "your-newsapi-key" } } }

{ "servers": { "scout-mcp": { "command": "python", "args": ["-m", "scout_mcp.mcp_server"], "cwd": "/path/to/scout-mcp/src", "env": { "NEWS_API_KEY": "your-newsapi-key" } } } }

pip install scout-mcp scout-mcp # starts STDIO server for MCP clients

Every response includes adata_quality_grade— a letter grade that lets agents instantly assess intelligence reliability:

result = scout_company("Stripe") if result["data_quality_grade"] in ("A+", "A"): # High confidence — safe to make decisions on this data proceed_with_analysis(result) elif result["data_quality_grade"] == "B": # Good but verify key claims proceed_with_caveats(result) else: # C/D/F — supplement with additional sources request_more_data(result)

Beyond the letter grade, every response includes aconfidence_breakdowndict showing per-source reliability:

{ "confidence": 0.86, "data_quality_grade": "A", "confidence_breakdown": { "duckduckgo": { "score": 0.60, "reason": "8 results found" }, "company_website": { "score": 0.90, "reason": "scraped stripe.com, 3 data points extracted" }, "wikipedia": { "score": 0.90, "reason": "page found, structured data extracted" }, "newsapi": { "score": 0.90, "reason": "6 articles found" }, "competitor_extraction": { "score": 0.80, "reason": "5 competitors identified" } } }

Different sources carry different weights in the overall confidence calculation:

Pass your API key via theX-Api-Keyheader:

curl -X POST /api/scout/company \ -H "X-Api-Key: your-key-here" \ -H "Content-Type: application/json" \ -d '{"name": "Stripe"}'

Response fields:name,domain,description,industry,founded,headquarters,employee_range,funding,tech_stack,social_profiles,recent_news,top_competitors,key_people,confidence,confidence_breakdown,data_quality_grade,data_freshness,sources_used,sources_failed

Response fields:market_name,query,depth,market_size,market_size_projections,cagr,key_players,trends,growth_drivers,risks,source_links,confidence,confidence_breakdown,data_quality_grade

Response fields:target,competitors(array of{name, domain, positioning, pricing, strengths, weaknesses, key_differentiator}),market_positioning_summary,confidence,confidence_breakdown,data_quality_grade

Response fields:topic,timeframe,sentiment({score, label}),trending_direction,key_developments,related_topics,social_buzz,confidence,confidence_breakdown,data_quality_grade

Response fields:name,category,description,pricing,ratings,features,ideal_for,alternatives,recent_updates,confidence,confidence_breakdown,data_quality_grade

Research a public figure.Requires Pro tier or higher.

Response fields:name,current_role,company,location,background_summary,social_profiles,recent_activity,notable_achievements,confidence,confidence_breakdown,data_quality_grade

{ "name": "Stripe", "domain": "stripe.com", "description": "Stripe, Inc. is an Irish-American multinational financial services and software as a service company...", "industry": "Software & Technology", "founded": "2010", "headquarters": "San Francisco", "top_competitors": ["PayPal", "Adyen", "Square", "Braintree", "Checkout.com"], "tech_stack": ["Next.js", "React"], "recent_news": [ {"headline": "Stripe launches AI billing features", "source": "TechCrunch", "date": "2026-04-08"}, {"headline": "Stripe revenue grows 30% in 2025", "source": "Bloomberg", "date": "2026-03-15"} ], "confidence": 0.86, "data_quality_grade": "A", "confidence_breakdown": { "duckduckgo": {"score": 0.60, "reason": "8 results found"}, "company_website": {"score": 0.90, "reason": "scraped stripe.com, 3 data points extracted"}, "wikipedia": {"score": 0.90, "reason": "page found, structured data extracted"}, "newsapi": {"score": 0.90, "reason": "6 articles found"}, "competitor_extraction": {"score": 0.80, "reason": "5 competitors identified"} }, "data_freshness": "2026-04-12T12:00:00Z" }

scout_trends("generative AI", timeframe="7d")

{ "topic": "generative AI", "timeframe": "7d", "sentiment": {"score": 0.72, "label": "positive"}, "trending_direction": "up", "key_developments": [ { "headline": "OpenAI releases GPT-5.2 with reasoning capabilities", "date": "2026-04-10", "impact_level": "high", "source": "The Verge" } ], "related_topics": ["Large Language Models", "AI Safety", "Enterprise AI"], "confidence": 0.83, "data_quality_grade": "A", "confidence_breakdown": { "duckduckgo_news": {"score": 0.90, "reason": "10 news articles found"}, "newsapi": {"score": 0.90, "reason": "4 articles found"}, "web_search": {"score": 0.50, "reason": "5 web results for context"} } }

{ "target": "Notion", "competitors": [ { "name": "Obsidian", "domain": "obsidian.md", "positioning": "Privacy-focused local-first knowledge base with Markdown", "strengths": ["Open source", "Free tier available", "Offline support"], "key_differentiator": "Local-first with plain Markdown files" }, { "name": "Coda", "positioning": "All-in-one doc with app-building capabilities", "strengths": ["AI-powered features"], "key_differentiator": "Document-as-app paradigm" }, { "name": "Logseq", "positioning": "Open-source outliner with bidirectional links" } ], "market_positioning_summary": "Found 7 competitors for Notion. Top alternatives: Obsidian, Coda, Logseq, Anytype, AppFlowy.", "confidence": 0.65, "data_quality_grade": "B", "confidence_breakdown": { "search_Notion alternative": {"score": 0.70, "reason": "8 results for 'Notion alternatives'"}, "search_Notion vs competi": {"score": 0.60, "reason": "6 results for 'Notion vs competitors'"}, "extraction_quality": {"score": 0.86, "reason": "7 competitors extracted and enriched"} } }

+------------------+ | AI Agent | | (Claude, etc.) | +--------+---------+ | STDIO / SSE / REST | +----------------+----------------+ | Scout MCP | | FastMCP 3.x + FastAPI REST | +----------------+----------------+ | Cache (24h TTL, in-memory) | | Auth (API key, tier limits) | | Rate Limiter (per-key, daily) | +----+--------+--------+----------+ | | | +--------+ +-----+--+ +--+--------+ | DuckDuckGo| | NewsAPI | | Wikipedia | | (free) | | (.org) | | (free) | +-----------+ +---------+ +-----------+ | | +--------+--------+ +-------+-------+ | Web Scraper | | Social Profile | | (httpx + BS4) | | Detection | +----------------+ +----------------+

DuckDuckGo's free API has rate limits. Scout MCP handles this with adaptive backoff:

Success : interval = 2s (base) Failure 1: interval = 4s Failure 2: interval = 8s Failure 3: interval = 16s Failure 4: interval = 32s Failure 5: interval = 48s (cap) Next success: interval resets to 2s

Each call also retries once before giving up. Monitor backoff status atGET /api/health.

The competitor extraction uses 6 regex pattern categories with 350+ stop words:
- VS patterns— "X vs Y" matching
- Comma/and-separated lists— "alternatives include X, Y, and Z"
- Numbered/bulleted lists— "1. Asana 2. Monday 3. ClickUp"
- Header patterns— "Asana -- project management tool"
- Contextual patterns— "like X" or "such as X"
- Title-cased names— capitalized product names near competitor context

Multi-word validators reject: article titles, verb-prefixed names, pronoun-prefixed names, role-suffixed names, probable person names, mega-corp parent names, and platform names.

Responses are cached for24 hoursin-memory (dict). Cache stats and clear endpoints:

# Check cache stats curl /api/cache/stats # Clear all cache curl -X POST /api/cache/clear

Rate limit info is included in every response's_metafield:

{ "_meta": { "tier": "free", "remaining": 47 } }

# Clone and install git clone https://github.com/your-org/scout-mcp.git cd scout-mcp/backend pip install -e ".[dev,server]" # Set up environment echo "NEWS_API_KEY=your-key" > .env # Run MCP server (STDIO for Claude Desktop) cd src && python -m scout_mcp.mcp_server # Run REST API server uvicorn server:app --host 0.0.0.0 --port 8001 --reload # Inspect with MCP Inspector fastmcp inspect src/scout_mcp/mcp_server.py

# Test the API curl -X POST http://localhost:8001/api/scout/company \ -H "Content-Type: application/json" \ -d '{"name": "OpenAI"}' # Check health + backoff status curl http://localhost:8001/api/health # List all tools curl http://localhost:8001/api/tools

FROM python:3.12-slim WORKDIR /app RUN apt-get update && apt-get install -y --no-install-recommends gcc libxml2-dev libxslt1-dev && rm -rf /var/lib/apt/lists/ COPY pyproject.toml . COPY src/ src/ RUN pip install --no-cache-dir . EXPOSE 8001 CMD ["python", "-c", "from scout_mcp.mcp_server import mcp; mcp.run(transport='sse', port=8001)"]

# Build and run docker build -t scout-mcp . docker run -p 8001:8001 -e NEWS_API_KEY=your-key scout-mcp

Built with FastMCP, httpx, BeautifulSoup4, Pydantic
Scout MCP v0.1.0*

🚀 Also check outProfitSpot MCP— cross-chain DeFi intelligence for AI agents. Risk-scored yields, Monte Carlo simulations, whale tracking across 86 chains.

SerpApi MCP Server for Google and other search engine results