Cutting wasted SaaS spend with AI agents that find unused seats, duplicate tools and owed SLA credits. Built @ Ramp Hackathon.
Preview
Features
- Paid vs. Used Dashboard: Spend, recoverable savings and overlapping tools across every vendor
- Agent Reviews: AI agents check each vendor's seat usage, plan tier and SLA uptime, streaming steps live
- Slack Reports: A full sweep posts its findings to Slack
- Email Drafts: One-click emails for seat cuts, consolidation, SLA claims and downgrades
- Spend Chat: Ask questions about vendors, seats and people
Tech Stack
- Frontend: React 19, Next.js 16 (App Router)
- Styling: Tailwind CSS v4, shadcn/ui, Motion, Recharts
- AI: Custom tool-calling agents on OpenAI, Anthropic or Gemini
- Language: TypeScript
Local Setup
- Clone the repository:
git clone https://github.com/maxtmiller/Lean.git
cd Lean
- Install dependencies:
npm install
- Configure environment variables:
cp .env.example .env.local
- Set
APP_MODE=prod to call the LLM live (leave it unset for demo mode)
- Set
LLM_PROVIDER and the matching OPENAI_API_KEY or ANTHROPIC_API_KEY
- Optionally add
SLACK_WEBHOOK_URL to post sweep reviews to Slack
- Run the development server:
npm run dev
- Open http://localhost:3000 to view the dashboard.
See DEVELOPMENT.md for scripts, demo recordings and the code layout.
Next Steps
- Real Data Sources: Connect a card feed and SSO logs instead of mock data
- Scheduled Sweeps: Run the agents weekly and alert before renewals
- Send from the App: Send drafted emails directly and track vendor replies
