Daloopa Starter Kit for Claude Code

A ready-to-go financial analysis toolkit that connects Claude Code to Daloopa's institutional-grade financial data. Built for hedge fund analysts (L/S equity, quant) who want AI-assisted fundamental research.

Produces investment deliverables: Research Notes (.docx), Excel Models (.xlsx), HTML Reports, and Pitch Decks (.pdf).

Prerequisites

• Claude Code — Install with npm install -g @anthropic-ai/claude-code
• Daloopa account — Sign up at daloopa.com
• Python 3.9+ — For infrastructure scripts (market data, charts, Excel/Word/PDF rendering)

Quick Start

# 1. Clone the repo
git clone https://github.com/daloopa/investing.git && cd investing

# 2. Install Python dependencies
pip3 install -r requirements.txt

# 3. Open in Claude Code
claude

# 4. Run the setup wizard
/setup

The /setup command will walk you through authenticating with Daloopa (OAuth opens in your browser), verifying the MCP connection, and running your first query.

That's it. On your first Daloopa tool call, OAuth will prompt you to log in via your browser. No API keys or .env files needed.

Available Commands

Building Block Skills (HTML Reports)

Investment Deliverables (.docx, .xlsx, .pdf)

All reports are saved to the reports/ directory. You can also just ask Claude anything about a company — the commands are shortcuts for common workflows.

Plugin

The 10 building block analysis skills are also available as a standalone Claude Code plugin that works in any project — no Python infrastructure needed, just a Daloopa account.

See daloopa/plugin or install from the Claude Code marketplace.

Data Access

MCP Server (Default — Interactive with Claude Code)

Two MCP servers are pre-configured in .mcp.json:

The data server provides 4 tools:

By default, authentication is via OAuth — a browser window opens on your first tool call. No API keys needed.

If you prefer API key auth for the MCP server (e.g., headless environments), update .mcp.json:

{
  "mcpServers": {
    "daloopa": {
      "type": "http",
      "url": "https://mcp.daloopa.com/server/mcp",
      "headers": {
        "x-api-key": "${DALOOPA_API_KEY}"
      }
    },
    "daloopa-docs": {
      "type": "http",
      "url": "https://docs.daloopa.com/mcp"
    }
  }
}

Then add your key to .env:

DALOOPA_API_KEY=your_api_key_here

Direct REST API (Programmatic — Python Scripts)

The same API key also works with the Daloopa REST API directly. The recipes/ directory contains Python scripts for headless automation, batch processing, or building custom pipelines.

All scripts use recipes/daloopa_client.py for authentication (Basic Auth with email + API key).

Setup for API access:

cp .env.example .env
# Edit .env with your credentials:
#   DALOOPA_EMAIL=you@example.com
#   DALOOPA_API_KEY=your_api_key_here

Then run any recipe:

python3 recipes/company_fundamentals.py AAPL
python3 recipes/document_search.py "AI revenue" --companies AAPL MSFT
python3 recipes/export_csv.py AAPL

The Claude Code skills auto-detect which access method is available and use whichever is configured. See .claude/skills/data-access.md for details.

Full API docs: docs.daloopa.com

Project Structure

├── .claude/
│   └── skills/                # Claude Code skill definitions
│       ├── data-access.md     # Shared data access reference
│       ├── design-system.md   # Formatting and styling conventions
│       ├── setup/             # /setup — interactive setup wizard
│       ├── earnings-review/   # /earnings-review — earnings analysis
│       ├── earnings-prep/     # /earnings-prep — pre-earnings preparation
│       ├── earnings-flash/    # /earnings-flash — rapid first-read flash
│       ├── tearsheet/         # /tearsheet — company one-pager
│       ├── industry/          # /industry — cross-company comp
│       ├── bull-bear/         # /bull-bear — scenario analysis
│       ├── guidance-tracker/  # /guidance-tracker — guidance vs actuals
│       ├── inflection/        # /inflection — acceleration/deceleration detection
│       ├── capital-allocation/# /capital-allocation — capital deployment
│       ├── dcf/               # /dcf — DCF valuation
│       ├── comps/             # /comps — trading comparables
│       ├── precedent-transactions/ # /precedent-transactions — M&A deal comps
│       ├── comp-sheet/        # /comp-sheet — industry comp Excel model
│       ├── supply-chain/      # /supply-chain — interactive supply chain dashboard
│       ├── unit-economics/    # /unit-economics — unit economics decomposition
│       ├── working-capital/   # /working-capital — cash conversion & earnings quality
│       ├── meta-skill/        # Internal: convert skills to MCP prompt functions
│       ├── ib-deck/           # /ib-deck — pitch deck builder
│       ├── research-note/     # /research-note — Word document output
│       ├── build-model/       # /build-model — Excel model output
│       ├── initiate/          # /initiate — both outputs
│       └── update/            # /update — refresh coverage
├── recipes/                   # Python scripts for direct API access
│   ├── daloopa_client.py      # Shared HTTP client with auth
│   ├── company_fundamentals.py
│   ├── document_search.py
│   ├── export_csv.py
│   ├── download_model.py
│   ├── industry_analysis.py
│   ├── taxonomy_comparison.py
│   ├── poll_for_updates.py
│   └── series_continuation.py
├── infra/                     # Infrastructure scripts (used by skills)
│   ├── market_data.py         # Market data fallback (yfinance/FRED)
│   ├── chart_generator.py     # Professional chart generation (6 types)
│   ├── projection_engine.py   # Forward financial projections
│   ├── excel_builder.py       # Multi-tab Excel model builder (single-company)
│   ├── comp_builder.py        # Multi-company comp sheet builder (8 tabs)
│   ├── docx_renderer.py       # Word document renderer
│   ├── pdf_renderer.py        # Markdown → styled PDF
│   ├── deck_renderer.py       # HTML deck → PDF
│   └── report_differ.py       # Context diff for updates
├── templates/
│   └── research_note.docx     # Word template (Jinja2 tags)
├── scripts/
│   ├── create_template.py     # Generate the Word template
│   ├── sync_plugin.sh         # Sync shared skills to plugin repo
│   └── docs_crawler.py        # Re-crawl Daloopa docs
├── daloopa_docs/              # API documentation (local copy)
├── reports/                   # Generated reports (gitignored)
│   ├── .charts/               # Generated charts
│   └── .tmp/                  # Context JSON files
├── .mcp.json                  # MCP server config
├── .env.example               # API key template
├── requirements.txt           # Python dependencies
├── CLAUDE.md                  # AI assistant instructions
└── README.md

Optional API Keys

Add to .env if desired. Without FRED, DCF calculations default to a 4.5% risk-free rate.

Refreshing Documentation

To re-crawl the Daloopa docs (e.g., after API updates):

python3 scripts/docs_crawler.py