Skip to content

OSOJDJD/deeplook

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

20 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DeepLook

Open-source MCP server for company research. 10 sources, structured output, under 15 seconds.

LLMs hallucinate financial data. DeepLook gives them real data instead — prices, financials, peers, news, technicals — from APIs, not from training data.

Python 3.10+ MIT License MCP Compatible


Quick start

Hosted (30 seconds):

Claude.ai → Settings → Connectors → Add MCP Server
URL: https://mcp.deeplook.dev/mcp

Then: "Use DeepLook to research NVIDIA"

Works with Claude, Cursor, Windsurf, or any MCP client.

Self-host:

git clone https://github.com/OSOJDJD/deeplook.git
cd deeplook
python3 -m venv venv && source venv/bin/activate
pip install -e .
cp .env.example .env   # add at least one LLM key
python -m deeplook.mcp_server --http --port 8819

Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "deeplook": {
      "command": "/full/path/to/deeplook/venv/bin/python",
      "args": ["-m", "deeplook.mcp_server"],
      "cwd": "/full/path/to/deeplook",
      "env": { "ANTHROPIC_API_KEY": "sk-ant-..." }
    }
  }
}

Replace ANTHROPIC_API_KEY with your preferred provider key: OPENAI_API_KEY, GEMINI_API_KEY, or DEEPSEEK_API_KEY.

CLI (no MCP):

python -m deeplook "NVIDIA"
python -m deeplook "Solana"
python -m deeplook "Anthropic"

Example output

DeepLook NVIDIA Report


Tools

Tool What it does Speed
deeplook_research Full report — financials, peers, news, technicals, catalysts ~15s
deeplook_lookup Quick snapshot — price, key metrics, headline ~3s

Supported entities

Works for public equities, crypto tokens, DeFi protocols, private companies, VC firms, exchanges, and foundations.


How it works

Company name
    ↓
Entity router → stock / crypto / private / VC / exchange / foundation / defunct
    ↓
10 parallel fetchers → yfinance, news, CoinGecko, SEC EDGAR, Wikipedia, ...
    ↓
Code layer extracts all numbers from APIs (not LLM-generated)
    ↓
LLM compress + judge → verdict, signals, catalysts
    ↓
Structured report (markdown + embedded JSON)

Data sources

yfinance · DuckDuckGo News · Wikipedia · YouTube · CoinGecko · RootData · DeFiLlama · SEC EDGAR · Finnhub · Company websites

API keys

At least one LLM provider required for self-host:

ANTHROPIC_API_KEY · OPENAI_API_KEY · GEMINI_API_KEY · DEEPSEEK_API_KEY

Optional: TAVILY_API_KEY · COINGECKO_API_KEY · ROOTDATA_SKILL_KEY

See .env.example for details.

Eval

58 companies tested (mega-cap, growth, crypto, pre-IPO, international):

Overall 3.78/5.0 · Risk detection 4.36/5.0 · Signal quality 3.94/5.0

Framework in /eval. Run it yourself, contribute ground truth.


Contributing

Found a bug? Report looks wrong? Have an idea? We'd love your help. See CONTRIBUTING.md to get started.

License

MIT


Built by OSOJDJD

About

No description, website, or topics provided.

Resources

License

Contributing

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages