PitchLense MCP

""" LV-Analysis MCP Tool for detailed startup analysis based on hackathon requirements. """ from typing import Dict, Any, List, Optional from fastmcp import FastMCP from ..core.base import BaseMCPTool class LVAnalysisAnalyzer: """Analyzer for LV-Analysis detailed startup analysis.""" def __init__(self, llm_client=None, perplexity_tool=None): self.llm_client = llm_client self.perplexity_tool = perplexity_tool def analyze(self, startup_text: str) -> Dict[str, Any]: """ Perform detailed LV-Analysis based on hackathon requirements. Args: startup_text: The startup description and details Returns: Detailed analysis in hackathon format """ # Get additional market information using Perplexity market_research = self._get_market_research(startup_text) market_info = market_research.get("research_text", "") market_sources = market_research.get("sources", []) # Combine startup text with market research enhanced_context = f""" Startup Information: {startup_text} Market Research Data: {market_info} Please analyze this startup and provide a detailed business note in the following format: """ prompt = self._build_analysis_prompt(enhanced_context) try: response = self.llm_client.predict( system_message="You are an expert startup business analyst specializing in detailed business note generation for investment evaluation. Maintain professional language and avoid inappropriate content. Focus strictly on business and investment analysis.", user_message=prompt ) analysis_result = response.get("response", "") # Extract structured data from the response structured_analysis = self._extract_structured_data(analysis_result, startup_text, market_info) # Add sources to the analysis result structured_analysis["sources"] = market_sources return structured_analysis except Exception as e: return { "error": f"Analysis failed: {str(e)}", "category_name": "LV-Analysis", "analysis_type": "detailed_business_note" } def _get_market_research(self, startup_text: str) -> Dict[str, Any]: """Get market research data using Perplexity.""" if not self.perplexity_tool: return { "research_text": "Market research unavailable: Perplexity tool not provided", "sources": [] } try: # Extract key terms for market research research_queries = [ "Indian FMCG market size 2024 healthy food trends", "millet based products market India growth statistics", "clean label food brands India market analysis", "women led startups India FMCG sector trends", "B2B food distribution India market opportunities" ] market_data = [] all_sources = [] for query in research_queries: try: result = self.perplexity_tool.search_perplexity(query) if result and "answer" in result: market_data.append(f"Query: {query}\nResult: {result['answer']}\n") # Collect sources sources = result.get("sources", []) if sources: all_sources.extend(sources) except Exception as e: print(f"Perplexity search failed for {query}: {e}") continue # Deduplicate sources by URL seen_urls = set() unique_sources = [] for source in all_sources: url = source.get("url") if url and url not in seen_urls: seen_urls.add(url) unique_sources.append(source) return { "research_text": "\n".join(market_data), "sources": unique_sources } except Exception as e: return { "research_text": f"Market research unavailable: {str(e)}", "sources": [] } def _build_analysis_prompt(self, context: str) -> str: """Build the analysis prompt for the LLM.""" return f""" Based on the following startup information and market research, provide a detailed business analysis in the exact format specified below: {context} Please provide a comprehensive analysis covering: 1. Industry and Market Size: - Industry Reports - Market Size, Others - Technology Reports - Food Innovation with AI, AI in Ops, AI in CPG 2. Revenue Streams (for each stream include): - Name of the Revenue Stream - Description (What is it, How does it work) - Target Audience (Who is paying) - Percentage Contribution (Share of total revenue) 3. Pricing Strategy: - Outline pricing models and tiers - Rationale behind pricing (market research, competitor analysis) 4. Unit Economics: - Customer Acquisition Cost (CAC) - Lifetime Value (LTV) - LTV:CAC Ratio 5. Recurring vs. One-Time Revenue: - Segregate revenue into recurring and one-time - Provide details for each 6. Payment Flow and Terms: - How payments are collected and processed - Payment frequency - Refund and cancellation policies 7. Scalability of Revenue Model: - How the revenue model will scale as business grows 8. Additional Revenue Opportunities: - Future revenue streams to explore 9. Competitor Analysis Framework: - Cover 2-3 competitors with details in table format: Category | Competitor 1 | Competitor 2 | Competitor 3 Company Name, Headquarters, Founding Year, Total Funding, Business Model, Revenue Streams, Target Market, etc. 10. Founders Profile: - Education, Work experience, Details of previously founded companies 11. Financials: - MRR, ARR, Burn, Runway, Gross Margin, CM1%, CM2%, CM3% 12. Facilities: - Office details, Plant details, Warehouses 13. Technology: - Tech stack writeup 14. Fundraiser: - Total funding details till date 15. Valuation: - Valuation rationale 16. Round Structure: - Terms, Pre-Money, Lead, Incoming Investors, Existing Investors 17. Key Problem Solved: - Clear problem statement and solution 18. Business Model: - Detailed business model explanation 19. Pipeline: - Sales Pipeline Value - Projected Growth Opportunities 20. Why Now: - Market Trends - Competitive Edge - Urgency/Opportunity 21. Financials: - Funding Ask - Structure - Valuation Cap and Floor - Current Commitments 22. Risks and Mitigation: - Identified Risks and Proposed Mitigation Strategies in table format Provide detailed, specific information for each section based on the startup data provided. """ def _extract_structured_data(self, analysis_result: str, startup_text: str, market_info: str) -> Dict[str, Any]: """Extract and structure the analysis data.""" return { "category_name": "LV-Analysis", "analysis_type": "detailed_business_note", "analysis_result": analysis_result, "startup_context": startup_text, "market_research": market_info, "sections": { "industry_market_size": self._extract_section(analysis_result, "Industry and Market Size"), "revenue_streams": self._extract_section(analysis_result, "Revenue Streams"), "pricing_strategy": self._extract_section(analysis_result, "Pricing Strategy"), "unit_economics": self._extract_section(analysis_result, "Unit Economics"), "recurring_vs_onetime": self._extract_section(analysis_result, "Recurring vs. One-Time Revenue"), "payment_flow": self._extract_section(analysis_result, "Payment Flow and Terms"), "scalability": self._extract_section(analysis_result, "Scalability of Revenue Model"), "additional_opportunities": self._extract_section(analysis_result, "Additional Revenue Opportunities"), "competitor_analysis": self._extract_section(analysis_result, "Competitor Analysis Framework"), "founders_profile": self._extract_section(analysis_result, "Founders Profile"), "financials": self._extract_section(analysis_result, "Financials"), "facilities": self._extract_section(analysis_result, "Facilities"), "technology": self._extract_section(analysis_result, "Technology"), "fundraiser": self._extract_section(analysis_result, "Fundraiser"), "valuation": self._extract_section(analysis_result, "Valuation"), "round_structure": self._extract_section(analysis_result, "Round Structure"), "key_problem": self._extract_section(analysis_result, "Key Problem Solved"), "business_model": self._extract_section(analysis_result, "Business Model"), "pipeline": self._extract_section(analysis_result, "Pipeline"), "why_now": self._extract_section(analysis_result, "Why Now"), "funding_ask": self._extract_section(analysis_result, "Funding Ask"), "risks_mitigation": self._extract_section(analysis_result, "Risks and Mitigation") } } def _extract_section(self, text: str, section_name: str) -> str: """Extract a specific section from the analysis text.""" try: lines = text.split('\n') section_start = None section_end = None for i, line in enumerate(lines): if section_name.lower() in line.lower(): section_start = i break if section_start is None: return f"Section '{section_name}' not found in analysis" # Find the next section or end of text for i in range(section_start + 1, len(lines)): if any(keyword in lines[i].lower() for keyword in [ "industry and market", "revenue streams", "pricing strategy", "unit economics", "recurring vs", "payment flow", "scalability", "additional revenue", "competitor analysis", "founders profile", "financials", "facilities", "technology", "fundraiser", "valuation", "round structure", "key problem", "business model", "pipeline", "why now", "funding ask", "risks and mitigation" ]) and i > section_start: section_end = i break if section_end is None: section_end = len(lines) return '\n'.join(lines[section_start:section_end]).strip() except Exception as e: return f"Error extracting section '{section_name}': {str(e)}"