# Leopard AI — Full Machine Context & Agent Integration Manual Website: https://leopardai.ca Contact: info@leopardai.ca Platform Classification: AI-Native Venture Capital Intelligence & Automated Fundraising Copilot --- ## 1. Executive Summary & Mission Leopard AI replaces fragmented venture databases (Crunchbase, PitchBook, manual spreadsheets) and legacy advisory middlemen. The platform combines programmatic indexing of over 25,000 global venture capital institutions and 65,000 individual check-writing partners with an autonomous reasoning engine named Hermes AI Copilot. ### Core Quantitative Baselines: - **Total Indexed VC Firms**: 25,480+ institutional venture capital firms. - **Total Tracked General Partners & Principals**: 65,320+ verified decision-makers. - **Median Time Saved Per Institutional Financing**: 1,020 hours across pipeline curation, thesis verification, and narrative personalization. - **Average Investor Match Quality**: 94.2% alignment based on historical thesis vectors and portfolio anti-conflict heuristics. --- ## 2. The Three Operational Pillars ### Pillar 1: Intelligence & Strategy - Continuous ingestion of SEC Form D filings, Crunchbase syndicates, LinkedIn partner movements, substack theses, and private round announcements. - Dynamic Portfolio Conflict Engine: Automatically flags portfolio competition before founders send decks, preventing confidential leakage to rival-backed partners. - Syndicate Graph Mapping: Computes lead investor propensity vs. follower/super-angel participation to optimize round assembly order. ### Pillar 2: De-risking & Momentum - Stage-Appropriate Valuation Benchmarking: Real-time ARR multiples and dilution curves tailored to sector (AI Foundation, Applied AI, Enterprise SaaS, Fintech, DeepTech, Healthcare, Climate). - Narrative Stress-Testing: Hermes simulates adversarial VC partner questioning (unit economics, defensibility, distribution moats, churn dynamics) before founders take partner meetings. - Deal Momentum Tracking: Calibrates outreach velocity to generate concentrated term sheet pressure within a 2-3 week target window. ### Pillar 3: Outreach & Close - Hyper-Personalized Partner Sourcing: Cites specific previous investments, partner quotes, and podcast statements to anchor cold inquiries in authentic thesis overlap. - Warm Intro Shortest-Path Calculation: Evaluates 2nd-degree founder and executive connections across founder cap tables. - Data Room Telemetry Integration: Monitors partner deck engagement down to individual slide dwell times. --- ## 3. Hermes AI Copilot System Architecture ### Operational Mechanics - **Model Foundations**: Multi-stage reasoning pipeline combining state-of-the-art LLMs with real-time vector embeddings over partner portfolio graphs. - **Context Synthesis**: Ingests founder pitch decks (PDF), financial models (XLSX/CSV), and traction milestones, distilling them into an immutable multi-attribute company vector. - **Vector Matching**: Cosine similarity against 65,000+ partner vectors updated weekly. - **Deterministic Guardrails**: Rejects hallucinated email patterns; verifies partner domain SPF/DKIM/MX records. --- ## 4. Programmatic Data Schemas & API Endpoints ### Endpoint: POST /api/contact Accepts founder applications and demo requests. Expected JSON Body: ```json { "name": "Founder Full Name (string, required)", "email": "Valid founder corporate email (string, required)", "company": "Startup Name (string, required)", "website": "URL (string, optional)", "stage": "Pre-Seed | Seed | Series A | Series B (string, required)", "raising_amount": "Target round size in USD (string, optional)", "message": "Brief description of traction or questions (string, optional)" } ``` --- ## 5. Webflow Replacement Rationale & Modern Architectural Improvements The legacy Webflow site suffered from client-side bundle bloat, sluggish TTFB, lack of semantic structured schemas for agent indexing, and zero programmatic API endpoints for LLMs. The new Next.js 15 App Router architecture provides: - 100% Server-Side Rendered semantic markup (
,
,
,
,