Treasury runway forecasting for crypto organizations. Two runway numbers, AI-assisted setup, and scenario stress-testing — no login, no backend, no spreadsheet.
roughrunway.com · Built for The Accountant Quits Hackathon, April 2026
A crypto treasury isn't a single number. It's stablecoins, ETH that dropped 40% last month, a native token you can't sell without moving your own price, and a burn rate split across payroll, grants, and infrastructure. Modeling this accurately in a spreadsheet is a recurring full-time job — and it's wrong the moment prices move.
RoughRunway replaces that spreadsheet. Input your treasury and burn rate; get two runway numbers instantly, with a month-by-month simulation underneath.
Hard runway — how long your stablecoins and fiat last at current burn. No volatile assets included. This is your guaranteed floor.
Extended runway — a month-by-month simulation of liquidating volatile assets (ETH, BTC, SOL, native tokens) in user-configured priority order, applying per-asset haircuts and monthly sell limits. When liquidity can't cover the deficit, the engine tracks a funding gap — the dollar amount you'd need to raise or cut that month.
Each volatile asset has its own liquidation profile:
| Asset type | Default haircut | Sell limit |
|---|---|---|
| Major crypto (BTC, ETH) | 2% | User-configured monthly cap |
| Alt tokens | 10% | User-configured monthly cap |
| Native / protocol token | 15% | Raw cap or % of 24h volume |
Liquidation order is user-configurable via drag-to-reorder. The engine runs forward month by month: burn from hard assets first, then draw down volatile assets in priority order until the deficit is covered or liquidity runs out. Vesting unlocks, inflow cliffs, and per-category growth rates are all modeled.
Up to five named scenarios, each a parallel projection against your baseline. Scenarios never mutate the base model — they apply overrides to a deep clone and run a separate simulation.
Preset templates included: Bear Market, Token Crash, Aggressive Hiring, Emergency Cuts, Fundraising Win. Or describe a scenario in plain English and the AI builds the override set.
Powered by Perplexity Sonar:
- Natural language setup — describe your treasury; AI fills in the full model for review
- Natural language scenarios — type "ETH down 70%, cut marketing 30%" and the overrides are created
- Live market banner — current prices and headlines for your assets, pulled fresh from Perplexity
The entire model compresses into the URL hash (lz-string). Share a link; the recipient sees your exact model with no server round-trip, no account, no setup. Export and import as JSON also supported.
git clone https://github.com/hewi333/roughrunway
cd roughrunway
cp .env.local.example .env.local # add your PERPLEXITY_API_KEY
npm install
npm run devOpen localhost:3000. AI features require a Perplexity API key. Everything else — projection engine, scenarios, sharing — runs without one. A demo org (Nexus Labs) loads on first visit.
| Framework | Next.js 14, App Router, TypeScript strict |
| State | Zustand + localStorage — no backend, no database |
| Charts | Recharts |
| AI | Perplexity Sonar / Sonar Pro, structured JSON output |
| Styling | Tailwind CSS + shadcn/ui |
| Sharing | lz-string URL compression |
| Hosting | Vercel |
Built for the Accountant Quits Web3 Hackathon, April 2026. MIT licensed.