Asphalt runway on a snow-covered alpine ridge at dawn, with snow-capped peaks glowing in warm light

Built for The Accountant Quits Hackathon

Know your runway.
Before you run out.

Crypto treasury runway modeling for Web3 orgs — stress-test scenarios and see exactly how long your money lasts.

Build modelRead the docs

4 clicks → sample model + sensitivity scenario

The Accountant QuitsPerplexity

The Platform

01

Runway modeling

Stablecoins, fiat, BTC/ETH, native tokens — each with tier-based liquidity haircuts and sell limits.

02

Scenario analysis

Model bear markets, hiring surges, grant cliffs, and one-off events side-by-side against your baseline.

03

Shareable links

Your entire model encodes into a URL. Send it to your accountant, board, or co-founder — no account needed.

Agentic

Tell Claude or ChatGPT.
Get a shareable model.

RoughRunway publishes /llms.txt and an OpenAPI spec so any AI assistant with web access reads your description, calls the encode API, and hands back a working model link — no forms, no copy-pasted JSON.

Works in

  • Claude Desktop via the roughrunway-mcp server — fully autonomous, model link comes back inline
  • Claude Code — uses its WebFetch tool natively, also fully autonomous
  • ChatGPT Custom GPT pointed at /openapi.json
  • Claude.ai / ChatGPT chat — web tools refuse to fetch model-built URLs, so the assistant parses your inputs and hands you a one-click link that redirects into the loaded dashboard

Discovery: /.well-known/ai-plugin.json · /llms.txt · /openapi.json

You, in any LLM chat

Build a runway model on roughrunway.com for our 15-person Web3 protocol post-$7M raise: $3M USDC + $500K USDT, $1M USD fiat, 300 ETH (Perplexity fills the price), 10M TAQ native tokens at $0.10 (15% haircut), $450K/mo burn across all eight categories (headcount $300K, infra $30K, legal $15K, marketing $35K, token incentives $25K, grants out $15K, office $8K, employee token grants $22K), $4K staking + $30K protocol revenue inflow. Project 18 months and return the dashboard link…

Assistant fetches (instant, no extra LLM call)

roughrunway.com/api/agent/encode?name=TAQ+Labs&stable=USDC:3000000,USDT:500000&fiat=USD:1000000&volatile=ETH:300:0:major,TAQ:10000000:0.10:native&burn=450000&team=15&months=18

Reply → click to open your model

Here's your runway model:

roughrunway.com/dashboard#model=N4Ig…