Getting Started

FundAdmin AI is a three-tier open-core platform for AI-powered fund administration.

Tiers

Tier	Description	License
T1 — Skills + CLI	62 Claude Code skills in skills/, orchestrated via /fund. MCP vault server in mcp-vault-server/.	MIT
T2 — Obsidian Vault	fund-admin-vault/ — FileClasses, dashboards, kanban boards, schema v1 frontmatter.	MIT
T3 — Webapp	Next.js 16 dashboard (apps/web), background worker (apps/worker), WebSocket sidecar (apps/ws), Postgres + Drizzle.	Commercial

T1 Quick Start (Skills + CLI)

cd fundadmin-ai
./install.sh       # installs skills to ~/.claude/skills/ and agents to ~/.claude/agents/

Then in Claude Code:

/fund review-lpa path/to/lpa.pdf

T3 Quick Start (Webapp)

cd fundadmin-ai
cp .env.example .env            # add ANTHROPIC_API_KEY; AUTH_SECRET is auto-generated
pnpm install
docker compose up -d postgres
pnpm db:migrate
pnpm seed                       # seeds 3 funds, 20 investors, 13 issues, 7 inbox items
pnpm dev                        # web :3000 + worker + ws :3002

Open http://localhost:3000, sign in with a magic link (delivered to the console in dev), and complete onboarding to create your first workspace and fund.

Useful Commands

# Run Python tests (T1 skill contracts, PDF generator, classifier, integration)
python3 -m unittest discover -s tests -v

# Typecheck the webapp
pnpm typecheck

# Run vitest unit tests
pnpm test

# Build docs
cd doc-site && npm run docs:build

Next Steps

• Architecture Overview — understand the three-tier system
• Commands Reference — all 62 /fund sub-skills
• LPA Safety Score — how the 5-agent review pipeline works
• API Reference — REST endpoints