The hedge fund knowledge graph
built for LLMs.

A complete, machine-readable encyclopedia of hedge funds and capital markets. 1,074 institutional-grade glossary terms, 12 strategies with expected vol & Sharpe profiles, 14 regulations across SEC / EU / ERISA, 20 parameterised calculators with worked examples, plus a live compute engine — all served as JSON, NDJSON, JSON-LD, or Markdown.

Quick start

Every endpoint is a static file or a Cloudflare Pages Function. No API key, no signup, no rate-limiting beyond Cloudflare's defaults. Try:

# Get the manifest of every endpoint, schema, and resource
curl https://hedgefund.wiki/api/v1/manifest | jq

# Fetch the canonical Sharpe Ratio definition with backlinks
curl 'https://hedgefund.wiki/api/v1/terms/sharpe-ratio?include=backlinks,graph' | jq

# Compute a Sharpe ratio
curl -X POST https://hedgefund.wiki/api/v1/compute/sharpe-ratio \
  -H 'content-type: application/json' \
  -d '{"R_p":0.12,"R_f":0.043,"sigma_p":0.10}' | jq

# Stream LLM-ready chunks for embeddings/RAG
curl https://hedgefund.wiki/api/v1/llms/chunks?kind=term

Architecture & operations

The API is served from generated artifacts plus Cloudflare Pages Functions. Source data lives in data/, scripts/build.py emits api-build/v1/*, and functions/api/v1/* handles query, graph, search, bulk, and compute behavior.

Engineering details, Mermaid diagrams, troubleshooting, and future improvement backlog live in ARCHITECTURE.md.

Discovery — start here if you're an LLM

Every entry-point an autonomous agent or RAG pipeline could need:

Endpoints

Compute engine

Every calculator in the corpus is also a callable endpoint. Inputs go in, outputs come out — including human-readable interpretation strings where appropriate.

POST /api/v1/compute/sharpe-ratio
Content-Type: application/json

{
  "R_p": 0.12,
  "R_f": 0.043,
  "sigma_p": 0.10
}

→ {
  "calculator": "sharpe-ratio",
  "name": "Sharpe Ratio",
  "inputs": { ... },
  "outputs": {
    "sharpe": 0.77,
    "interpretation": "Adequate."
  },
  "formula": "Sharpe = (R_p - R_f) / σ_p",
  "references": [{"title": "Mutual Fund Performance", "author": "William F. Sharpe", "year": 1966}]
}

Available calculators include: sharpe-ratio, sortino-ratio, calmar-ratio, information-ratio, treynor-ratio, jensens-alpha, var-historical, var-parametric, expected-shortfall, max-drawdown, kelly-criterion, black-scholes, implied-volatility, bond-duration, convexity, beta-capm, annualization, fund-of-funds-fees, high-water-mark-perf-fee, drawdown-recovery.

Content negotiation

Append ?format= or send an Accept header:

Schemas

Every entity has a JSON Schema (Draft 2020-12). Use them to typecheck or generate code:

• term.schema.json
• strategy.schema.json
• regulation.schema.json
• calculator.schema.json
• category.schema.json
• convention.schema.json
• source.schema.json
• person.schema.json
• fund.schema.json

Worked examples

Black-Scholes with Greeks

curl 'https://hedgefund.wiki/api/v1/compute/black-scholes?S=100&K=100&T=0.25&r=0.043&sigma=0.20&type=1'

Term graph 2-hop neighborhood

curl 'https://hedgefund.wiki/api/v1/graph/sharpe-ratio?depth=2'

Bulk fetch

curl -X POST https://hedgefund.wiki/api/v1/terms/_bulk \
  -H 'content-type: application/json' \
  -d '{"slugs":["sharpe-ratio","alpha","beta"],"fields":"slug,term,definition"}'

RAG ingestion

# 1,120 pre-chunked, embedding-ready text docs
curl https://hedgefund.wiki/api/v1/llms/chunks > corpus.ndjson

Architecture & operations

Maintainers and autonomous agents should read the architecture runbook before changing source data, generated artifacts, or Pages Functions. It covers the data build pipeline, Cloudflare Pages request path, local server harness, troubleshooting matrix, and next-iteration gaps.

License & attribution

All data is published under CC-BY-4.0. Attribution: hedgefund.wiki. Reuse for training, RAG pipelines, analytics, and downstream products is encouraged.