Fear & Greed

--

Loading...

BTC Dominance

--

Loading...

DeFi TVL

--

Loading...

Backend

● Checking...

Port 8888

● Daemon Status

Waiting for daemon heartbeat...

Quick Actions

📈 Watchlist

View All →

📄 Recent Reports

View All →

☑ Open Tasks

View All →

⚡ Recent Projects

View All →

💡 Auto-Discovery

Start backend to see auto-discovered opportunities

Mission

Build an AI-powered fundamental research engine that discovers asymmetric opportunities in crypto and equities before the market prices them in.

What We Do

Systematic fundamental analysis — quantitative scoring, on-chain financial statements, multi-factor models
24/7 autonomous scanning — the agent never sleeps, surfaces opportunities ranked by asymmetry (risk/reward × conviction)
Institutional-grade risk framework — portfolio-level VaR, correlation management, liquidity analysis

Why It Matters

Most funds rely on human analysts who miss opportunities, sleep, and have biases
We combine Jumparo's trading edge (mean reversion, narrative front-running, 5-lens filter) with AI that scales it to thousands of assets simultaneously
The research engine IS the edge — proprietary data, proprietary scoring, proprietary signals

Operating Principles

Every signal must be backtested before it goes live — no gut feelings at scale
Risk management is non-negotiable — 1% risk per position, correlation limits, counterparty diversification
Quality over quantity — A+ setups only, scored 18+/25 minimum
The agent learns — every alert tracked to outcome, weights adjusted, dead signals killed
Data advantage compounds — proprietary datasets widen the moat over time

Scale Trajectory

NOW $10-25K personal capital, building the prototype

PHASE 1-2 Institutional-grade data + quant models, ready for external capital

PHASE 3-5 Managing billions with full autonomy, multi-user, API platform

🔍

Project	Sector	Status	Source	Added

🔍

This Week

0

This Month

0

Positive

0

Negative

0

Intelligence Sources

𝕏 CryptoRank VCs 💎 DefiLlama Raises 📈 CryptoRank Rounds 🌱 RootData

Tracked Raises

0

Total Raised

$0M

Tier 1 Backed

0

100x Candidates

0

🔍

💡 Auto-Discovery Scanner

Deep Value Coins

Click Scan to find undervalued gems

Undervalued DeFi

Click Scan to find undervalued DeFi protocols

Total Value

$0.00

Total Cost

$0.00

Unrealized P&L

$0.00

Positions

0

Cash

$0.00

Holdings

No positions yet. Click "+ Add Position" to track your first holding.

Allocation Breakdown

Add positions to see allocation breakdown.

⚠ Risk Dashboard

Click "Analyze Risk" to calculate VaR, correlations, and position sizing.
Requires backend server running on port 8888.

🎯 Position Sizer

Capital ($)

Risk Per Trade (%)

Entry Price ($)

Stop Loss ($)

Market Regime

--

Load data to detect regime

Tracked Entities

0

Across all venues

Active Signals

0

Last 24h

Venues Connected

2

Hyperliquid + Dexscreener (Arkham soon)

Quick Access

Recent Activity

No recent on-chain activity tracked yet. Start by scanning wallets on Hyperliquid.

On-Chain / 🐋 Hyperliquid

🔍

Not loaded

BTC Regime

--

Score: --/5

Active Playbooks

--

Filtered by current regime

Open Setups

0

Watching for entry

Avg Confluence

--

Across active setups

Loading indicators...

Regime Rules

Risk-On

BTC funding positive, OI rising, dominance stable. Momentum + breakout playbooks active.

Neutral

Mixed signals. Mean reversion playbooks active. Reduced sizing.

Risk-Off

BTC funding negative, OI declining. Defensive only — shorts + hedges.

Senpi Rule: Mean reversion fails on perps without regime filter. Always check regime before activating a playbook.

Getting Started

🏠 Overview

🎯 Research Pipeline

🏗 Architecture

Agents

💬 Pamela (AI)

🐋 Hyperliquid Agent

⏰ Scheduled Scans

🚀 Agent Roadmap

Reference

⭐ Scoring System

🔗 APIs & Data Sources

🤖 AI Models

Updates

📜 Changelog

📅 Progress

Operations

🎯 About

☑ Tasks

🗺 Roadmap

Overview

Fund HQ is an AI-powered fundamental research engine that discovers asymmetric opportunities in crypto and equities before the market prices them in.

What is Fund HQ?

A single-page application for managing the full research lifecycle — from discovery to tracking. It combines live market data, automated scoring, AI-powered research, and on-chain whale intelligence into one dashboard.

Key Features

🔎

Market Data
Search crypto & stocks with live fundamentals

⭐

Auto Scoring
5-star quant ratings from live data

🤖

AI Research
Claude-powered deep analysis with web search

🐋

Whale Tracking
Hyperliquid on-chain wallet intelligence

🔔

Smart Alerts
Scheduled scans push signals to your feed

💬

Pamela Agent
20-tool AI assistant for fund ops

Pages

Page	Section	Description
Dashboard	Overview	Stats, quick actions, watchlist, reports, tasks at a glance
About	Overview	Mission, operating principles, scale trajectory
Project Dump	Pipeline	Quick capture for project ideas
Watchlist	Pipeline	Tracked assets with scores, detail view, Firebase sync
Alerts	Pipeline	Signals from scheduled scans + Pamela
On-Chain Intel	On-Chain	Smart money tracking — Hyperliquid, Entity Tracker, Flow Monitor, Signals
Market Data	Research	Live search, CoinGecko + FMP fundamentals
Templates	Research	Crypto (18s), Stock (14s), Quick Thesis (5s)
Reports	Research	Investment memos, research reports, Substack drafts
Technical	Technical	Market regime, setups, playbooks, confluence scoring, trade log
Tasks	Operations	Kanban board with milestones, priority, assignee
Roadmap	Operations	Tasks grouped by milestone with progress bars
Progress	Operations	Daily work log, timeline, streak tracking

Research Pipeline

A 5-step workflow from discovery to tracking — designed to surface asymmetric setups systematically.

1

Discover

Go to Market Data and search for any crypto or stock. Each result shows 3 quick actions:

+ Watch — add to watchlist for live tracking
+ Dump — save to Project Dump for later
AI Report — instantly generate a research report

Crypto uses CoinGecko (free, no key). Stocks need a free FMP key in Settings.

2

Score

Every asset gets a 5-star rating computed from live fundamentals:

Crypto: Momentum, Liquidity, Value (dilution), Market Position, Recovery
Stocks: Valuation (P/E), Profitability (ROE), Health (D/E), Momentum, Yield

AI research can add a ±0.5★ qualitative adjustment based on catalysts and sentiment.

3

Research

Pick a template in Templates:

Crypto Research — 18 sections (tokenomics, team, moats, whales, etc.)
Stock Research — 14 sections (business model, industry TAM, financials, capital allocation, valuation with DCF sensitivity, ownership, ESG, etc.)
Quick Thesis — 5 sections for fast screening

Fill manually or click AI to let Claude research it. Options: "Run All with AI" or "Fill Empty Only".

4

Report

Click Save as Report to compile sections into a formatted report. Reports appear in the Reports page where you can:

Set a Rating (Strong Buy, Buy, Hold, Sell, Avoid)
Set Conviction (High, Medium, Low)
Edit content or re-run AI with a different model

All reports sync in real-time between users via Firebase.

5

Track

Your Watchlist tracks live prices, 24h changes, key metrics, and scores. Click any card for the full detail view with score breakdown.

Hit Refresh All to update prices and re-compute scores for everything at once.

Architecture

How the system is built — stack, data flow, and deployment.

Tech Stack

Layer	Technology	Details
Frontend	Vanilla JS	Single HTML file, no build tools, dark theme
Database	Firebase RTDB	Real-time sync between users, localStorage fallback
Hosting	GitHub Pages	Free, auto-deploys from `jumparo1/JumpTools`
AI	Claude API	Haiku/Sonnet/Opus via tool_use, web search
On-chain	Hyperliquid API	Free, no auth — positions, trades, wallet data
Automation	Claude Code crons	3 scheduled tasks push alerts to Firebase

Data Flow

User → Fund HQ (Browser)
         ↓                    ↓                    ↓
   CoinGecko/FMP      Claude API         Hyperliquid API
   (market data)      (AI research)      (on-chain data)
         ↓                    ↓                    ↓
         →→→ Firebase RTDB ←←←
                  ↑
         Python Agent (cron / WebSocket)
         Claude Code scheduled tasks

Deploy Flow

cp fund-hq.html ~/deploy/
cd ~/deploy && git add . && git commit && git push
# GitHub Pages auto-builds in ~30s

API keys stored in browser localStorage (per-domain). Never committed to GitHub.

Pamela — AI Agent

20-tool AI assistant powered by Claude Haiku. Multi-turn conversations with up to 8 tool-use turns per query.

How to Use

Click the chat bubble (bottom-right corner) to open Pamela. Type naturally — she'll figure out which tools to use.

Example Commands

price btccrypto/stock price

watch ethadd to watchlist

research solAI research report

who's long ETH?whale positions

scan wallet 0x...profile wallet

whale activity on SOLlarge positions

what wallets are we tracking?saved wallets

create task: review TAO thesisnew task

All 20 Tools

Tool	Description
`search_crypto`	Search CoinGecko for crypto assets
`get_crypto_data`	Get detailed crypto fundamentals
`search_stocks`	Search FMP for stock tickers
`get_stock_data`	Get stock price, financials, ratios
`create_item`	Create a task, project, or dump entry
`update_item`	Update any item's fields
`delete_item`	Remove an item from the database
`find_items`	Query items by type and filters
`search_all`	Full-text search across everything
`add_to_watchlist`	Add asset to watchlist with scoring
`get_watchlist`	Get current watchlist with prices
`navigate_to_page`	Switch to any Fund HQ page
`get_fund_stats`	Dashboard stats (counts, streaks)
`save_research`	Save completed research as report
`get_reports`	List all saved research reports
`get_report_content`	Read full report content
`create_alert`	Push alert to Alerts page
`hl_scan_wallet`	Scan Hyperliquid wallet (positions, PnL)
`hl_whale_positions`	Find whale positions on a coin
`hl_saved_wallets`	Get list of tracked wallets

Hyperliquid Agent

On-chain whale tracker for Hyperliquid perpetual DEX. Finds profitable wallets and monitors whale activity.

Frontend (Fund HQ Page)

Three tabs in the Hyperliquid page:

Tab	What it does
Wallet Scanner	Paste any 0x address — see account value, open positions, PnL, win rate, trade history
Whale Discovery	Scans top coins via Hyperliquid API, finds $500K+ positions, sortable by size/PnL/leverage
Saved Wallets	Label and track wallets with live position updates, synced to Firebase

Python Agent

Backend agent for persistent monitoring. Located at agents/hyperliquid/agent.py.

python agent.py24/7 monitoring + WebSocket

python agent.py discoverfind whales on watched coins

python agent.py profile 0x...grade wallet A-F

python agent.py scancheck all saved wallets

Wallet Scoring

Wallets are graded A-F based on:

Win rate — percentage of profitable trades
Profit factor — gross profit / gross loss ratio
Leverage risk — penalty for consistently high leverage

Alerts push to Firebase when whales open $1M+ positions, close $500K+ positions, or when unrealized PnL swings 20%+.

API Reference

POST https://api.hyperliquid.xyz/info

// Get wallet positions (weight: 2)
{"type": "clearinghouseState", "user": "0x..."}

// Get trade history (weight: 20)
{"type": "userFills", "user": "0x..."}

// Get recent trades with addresses (weight: 20)
{"type": "recentTrades", "coin": "BTC"}

// Get market meta (OI, funding, volume)
{"type": "metaAndAssetCtxs"}

Scheduled Scans

3 automated Claude Code cron jobs that run on schedule and push alerts to Firebase.

Task	Schedule	What it does
Daily Market Scan	8am weekdays	Top movers, unusual volume, narrative shifts, asymmetric setups
Watchlist Refresh	8pm daily	Price changes, score shifts, breakout/breakdown signals
Weekly Deep Dive	Sunday 10am	Sector rotation, new narratives, top 3 opportunities with full scoring

All alerts appear in the Fund HQ Alerts page in real-time via Firebase REST API.

Agent Roadmap

Building toward multiple autonomous agents managing $1B AUM.

✓

Pamela — AI research + fund ops

20 tools, multi-turn conversations, web search, all fund operations

✓

Hyperliquid Agent — whale tracking + wallet profiling

Wallet scanner, whale discovery, A-F grading, WebSocket monitoring, Firebase alerts

✓

Market Scanner — scheduled scans + alerts

3 cron jobs: daily scan, watchlist refresh, weekly deep dive

☐

Risk Agent — portfolio risk management

VaR calculations, correlation analysis, position sizing, portfolio construction

☐

Execution Agent — automated trading

Signal-based trade execution, order management, slippage control

☐

Macro Agent — macro intelligence

Fed policy, rates, liquidity conditions, macro regime classification

Scoring System

Automated quantitative ratings computed from live market data.

Crypto Score (5 Stars)

Factor	What it measures
Momentum	24h and 7d price change strength
Liquidity	24h volume relative to market cap
Value	Market cap vs FDV dilution ratio
Market Position	Market cap rank and dominance
Recovery	Distance from all-time high

Stock Score (5 Stars)

Factor	What it measures
Valuation	P/E ratio vs sector average
Profitability	Return on equity (ROE)
Health	Debt-to-equity ratio
Momentum	Position within 52-week range
Yield	Dividend yield attractiveness

AI Overlay

When AI research runs, it can add a ±0.5★ qualitative adjustment based on catalysts, sentiment, and narrative strength.

5-Lens Filter (/25)

Advanced scoring for asymmetric setups. Each lens scores 1-5:

Catalyst — upcoming events, partnerships, launches
Valuation — relative to peers and intrinsic value
Momentum — price trend, volume confirmation
Risk:Reward — asymmetry of the setup
Narrative — market attention, story strength

APIs & Data Sources

External services powering Fund HQ's data layer.

API Keys

Service	Key Required?	Data Provided
CoinGecko	No (free)	Crypto prices, MCap, FDV, volume, supply, ATH, categories
FMP	Yes (free key)	Stock price, P/E, ROE, D/E, dividend yield, beta, sector
Anthropic	Yes (get key)	AI research, Pamela chat, web search
Hyperliquid	No (free)	Wallet positions, trades, OI, funding rates
DefiLlama	No (free)	TVL data for DeFi protocols

All keys stored in Settings → browser localStorage (per-domain). Never committed to GitHub.

Endpoint Notes

FMP: Use /stable/ endpoints (not v3). Field names: marketCap not mktCap.
Hyperliquid: All /info POST requests. Weight-based rate limiting (2-20 per call).
Claude: Direct browser access with anthropic-dangerous-direct-browser-access header.

AI Models

Claude model selection guide for different use cases.

Model	Speed	Cost	Best For
Haiku	Fastest	~$0.005/query	Quick lookups, Pamela chat, basic research
Sonnet	Medium	~$0.03/query	Balanced research, moderate analysis
Opus	Slowest	~$0.15/query	Deep fundamental analysis, complex research

Select model in the research workspace or report editor. Pamela defaults to Haiku for speed.

Changelog

Release history and feature updates.

March 20, 2026

Watchlist Split by Type — Crypto/Stocks tabs now filter the watchlist grid to show only relevant assets with per-type empty states
Watchlist Persistence Fix — Firebase listener race condition fixed with merge-based approach + dual-write (Firebase + localStorage)
Sparkline Charts — inline SVG 7-day sparklines on watchlist cards from CoinGecko, green/red based on trend
TradingView Chart — embedded advanced chart widget in watchlist detail view with dark theme
Add to Watchlist Button — prominent purple "+ Add" button on search results for one-click watchlist add
Stock Scoring Accuracy — replaced fake 2.5 defaults with N/A, hidden conviction badge when no real data
Auto-Enrich Stocks — every new stock auto-fetches financials, enriched metrics persist in Firebase
FMP API Fix — premium endpoint graceful degradation, free tier compatibility
Frontend-Backend Integration — wired all 5 backend phases into frontend: macro intelligence, auto-discovery, risk dashboard, performance analytics, smart alerts
Report Table UI — pipe-delimited table renderer with color-coded scores, styled card borders, accent headers, hover effects
Data-Driven Research Template — valuation scoreboard (10 dimensions /100), non-crypto comparison, strict data rules (every claim needs API data)
FET/ASI Research Report — 17-section deep research with perp positioning, whale analysis (6.5/10), valuation scoreboard (82/100)
FET Substack Article — 15-section newsletter with TOC, warm intro, clean text format
Report Timestamps — hours and minutes now shown alongside date on all report cards
Alerts Refreshed — stale test alerts removed, fresh data-driven alerts with live CoinGecko + Fear & Greed data
Docs Sidebar — Progress moved under Updates section (next to Changelog)
24/7 Agent System Planned — daemon architecture with 3 autonomous workers, agent memory, learning loops, launchd auto-start

March 19, 2026

Phases 1-5 Backend — FastAPI server with data feeds, quant models, risk engine, learning loop, autonomy layer
4 Background Loops — prices (2min), full data (10min), smart alerts (5min), auto-research (15min)
Mac Mini Infrastructure — SSH keys, repo clones, .env API key storage, launch.json config
VC Feed Scan — 4th scheduled task: daily DefiLlama raises fetch to Firebase

March 18, 2026

Auto Market Regime — live regime detection from 5 indicators: BTC funding, OI trend, dominance, Fear & Greed, BTC vs 200D SMA
get_market_regime Tool — Trader agent tool auto-populates regime for confluence scoring
VC Due Diligence Template — 18-section professional template with risk scoring, return modeling, 100x filter
VC Watchlist Page — track fundraising rounds from CryptoRank, DefiLlama, RootData with status tracking

March 17, 2026

5-Agent System Complete — Router, Analyst, Sentinel, Trader, VC agents with dedicated tools and system prompts
Sidebar Restructured — 4 agent sections with colored badges (Sentinel, Analyst, Trader, VC)
500-Wallet Token Intel — expanded whale scan from 30 to 500 wallets with concurrent batches
Wallet Hunter — systematic leaderboard scanning, wallet scoring (0-100), auto-classification
Arkham Intelligence — wallet identity labels, token holders, Alpha Summary insights
Whale Quality Scoring — composite score engine: win rate, PnL consistency, leverage analysis, auto-grade A+ to C
Moralis Integration — top EVM token holders with supply concentration analysis
Dexscreener Integration — token search, trending pairs, new listings
Intelligence OS Restructure — pages reduced from 17 to 11, templates merged into Reports, Signal/Flow merged

March 16, 2026

3-Division Architecture — restructured entire platform into On-Chain Intel, Research, and Technical Analysis divisions
On-Chain Intel — new top-level page with Overview dashboard, Hyperliquid (preserved), Entity Tracker, Flow Monitor, Signal Feed tabs
Technical Analysis — new page with Market Regime (risk-on/neutral/risk-off rules), Setup Scanner, Playbooks (3 strategies + backlog), Confluence scoring (4-factor), Trade Log
Perp Positioning — new Hyperliquid tab showing all 190 active perps with OI, funding, volume, premium
Report Trash — soft-delete with 30-day retention, restore, permanent delete, auto-purge
Stock template upgraded — expanded from 10 to 14 sections with institutional research rigor (Porter's Five Forces, DCF, ESG)
Hyperliquid Agent v2 — rewritten with Market Regime detection, Pattern Detector, Copy-Trade Signals, enhanced wallet scoring
Sidebar restructured — On-Chain Intel, Research, Technical, Operations sections
Logo navigation — click logo returns to Dashboard
Entity Tracker — full CRUD entity management, import from HL Saved Wallets, live position scanning, grade/type filtering
Setup Scanner — trade setup cards with entry/stop/target, auto R:R, playbook linkage, status cycling
Catalyst Tracker — event tracking with 8 types, impact/time filters, summary cards, Firebase sync
Confluence Scoring — 4 weighted sliders (On-Chain 25%, Fundamental 25%, TA 30%, Regime 20%), position sizing recommendations
Signal Feed — on-chain signal aggregation with type/conviction filters, status tracking, outcome win/loss tracking
Flow Monitor — aggregate entity flow scanning, accumulation/distribution detection, net flow by asset
Trade Log — full trade logging with P&L, R:R, playbook linkage, win rate, best playbook stats
Agent scaffolding — Sentinel, Analyst, Quant Python agents + skills files (ENTITY-INVESTIGATION, TRADE-PLAN)

March 15, 2026

Hyperliquid Whale Tracker — new page with Wallet Scanner, Whale Discovery, and Saved Wallets
Python Agent — persistent monitoring, wallet profiling (Grade A-F), WebSocket real-time trade alerts
Pamela upgraded to 20 tools — 3 new Hyperliquid tools: scan wallets, find whale positions, check tracked wallets
Firebase sync — saved wallets + discovered whales shared between frontend and Python agent
Docs page — GitBook-style documentation hub with sidebar navigation

March 13, 2026

Admin auth — login system with SHA-256 hashed credentials, view-only mode for visitors
Substack publishing — Reports split into Research/Substack tabs, 13-section article template
Report template — improved Catalysts & FUD section for research quality

March 12, 2026

Pamela AI Agent — upgraded from regex bot to Claude tool_use with 17 tools, multi-turn (max 8 turns)
Alerts page — severity-colored cards, expand/dismiss, browser notifications
Scheduled tasks — 3 cron jobs write alerts to Firebase
About page — fund mission, operating principles, scale trajectory
Knowledge base — 4 SKILL.md files: Mission, Trading Edge, Asymmetric Setups, Fundamental Analysis
Web search — AI research pulls live data and cites sources
5-phase roadmap — 35 items from Prototype to Full Autonomy

March 11, 2026

Initial launch — Dashboard, Watchlist, Market Data, Templates, Reports, Tasks, Progress
Scoring system — automatic 5-star ratings from live fundamentals (crypto + stocks)
GitHub Pages — migrated from Netlify (free, no bandwidth limits)
Firebase — all data syncs in real-time, localStorage fallback
AI Research — Claude API direct from browser, model selector (Haiku/Sonnet/Opus)
Market Data — CoinGecko (crypto) + FMP (stocks) search + fundamentals

Real-time Sync

Status: Connected

Your name:

Both users are synced in real-time via Firebase. Changes appear on the other screen within seconds.

Backup

Download a local JSON backup of all fund data.

Backend Server

FastAPI server for valuation, risk, performance analytics, and auto-discovery.

API Keys

Keys are stored locally in your browser. Never sent to our servers.

Free — No Key Required

CoinGecko (crypto prices) Hyperliquid (perps, wallets) Dexscreener (DEX pairs)

AI

Anthropic Paid

AI research reports, Pamela agent. console.anthropic.com

Research

Financial Modeling Prep Free

Stock data, fundamentals, financials (250 req/day). financialmodelingprep.com

On-Chain Intelligence

Arkham Intelligence Apply

Wallet identity labels, entity tracking, on-chain transfers, token holders. arkm.com

Moralis Free

Token Intel: top holders, % supply, whale categories (40K CU/day). moralis.com

Birdeye Free

Token Intel: top Solana DEX traders, buy/sell volume. birdeye.so

CoinGlass Free

Whale tracker, liquidations, OI data. coinglass.com

Danger Zone

This deletes ALL data for BOTH users. Cannot be undone.

Fund HQ Setup

Dashboard

● Daemon Status

📈 Watchlist

📄 Recent Reports

☑ Open Tasks

⚡ Recent Projects

💡 Auto-Discovery

About — Fund HQ

Mission

What We Do

Why It Matters

Operating Principles

Scale Trajectory

Deal Flow

Project Dump

Watchlist

Signals & Alerts

Research Notes

Research Templates

Reports

📅 Catalysts

💰 VC Watchlist

Intelligence Sources

Tasks

Roadmap

Progress

Market Data

💡 Auto-Discovery Scanner

Deep Value Coins

Undervalued DeFi

💰 Portfolio

Holdings

Allocation Breakdown

⚠ Risk Dashboard

🎯 Position Sizer

🔗 Intelligence

Quick Access

Recent Activity

Entity Flow Activity

📊 Trading

Regime Rules

Confluence Scorer

Scoring Guide

Performance by Setup

Score Calibration

Overview

What is Fund HQ?

Key Features

Pages

Research Pipeline

Discover

Score

Research

Report

Track

Architecture

Tech Stack

Data Flow

Deploy Flow

Pamela — AI Agent

How to Use

Example Commands

All 20 Tools

Hyperliquid Agent

Frontend (Fund HQ Page)

Python Agent

Wallet Scoring

API Reference

Scheduled Scans

Agent Roadmap

Scoring System

Crypto Score (5 Stars)

Stock Score (5 Stars)

AI Overlay

5-Lens Filter (/25)

APIs & Data Sources

API Keys

Endpoint Notes

AI Models