Know if your AI is making or losing money.
// theknowai = unit_economics(your_costs, your_price) → verdict
Every AI request has a real cost that flat SaaS pricing ignores. See your true gross margin, cost per action, and break-even runway — with a clear verdict in 30 seconds. No spreadsheets, no consultants.
no charge · 7 days · cancel anytime
The problem
Flat pricing was built for software that costs nothing to run
AI isn't that software. Tokens are a real, variable cost of goods sold — and that's exactly where the money leaks.
Every action burns money
Classic SaaS costs almost nothing per request. Each AI action spends tokens — your cost scales with usage, not just signups.
Flat pricing bleeds margin
A power user can cost 5× a light user to serve. One flat price overcharges some and quietly loses money on the rest.
You find out from the invoice
By the time the model bill lands, the price is already live. The margin damage is done — and it's hard to walk back.
How it works
From raw costs to the right price, in one place
Three steps: connect your costs, get a defensible price, and ship — without rebuilding a spreadsheet every time a model price changes.
Map your costs
Add LLM calls, infra, APIs, and labor — or let AI scan your product and surface the costs you forgot.
See your real margin
We compute cost per action and cost per user from live model prices, then Claude Opus recommends the pricing model that fits.
Simulate & ship
Project 12 months, find your break-even month, and generate a Stripe-ready pricing page.
Free tool · no signup
The 30-second margin check
Pick your LLM, enter usage and price, see whether you're profitable per user — or quietly funding your customers.
Pick your model
Claude, GPT, Gemini — live token prices from 2,000+ models.
Enter usage & price
Tokens per request, requests per user, your monthly price.
See your real margin
Gross margin, break-even price, dominant cost source.
See it in action
What you actually get inside
Real product views — margin simulator, AI pricing engine, and agent run simulator. Drop in your numbers, see your real economics.
12-month projections you can drag
See break-even month, gross margin, and MRR at the end of year one. Drag growth, churn, price, and actions/user sliders — chart updates live. Spot the month you stop losing money before you launch.
- Cost-at-scale: $9.38/user at 1K users, full breakdown
- Breakeven detection across optimistic / realistic / conservative
- Persists your inputs — refresh keeps your scenario
MRR at M12
$48.2k
Gross margin
68.4%
Break-even
Month 6
12-month gross margin
in %Cost at 1K users
$9.38/user/mo
$5.60 fixed + $3.78 actions
Claude Opus picks your pricing model
Answer six questions about your product, market, and usage. Claude Opus recommends flat-rate, usage-based, credit-based, or hybrid — with specific tier prices and the reasoning behind each.
- Specific dollar amounts per tier — not vague advice
- Flags hidden traps (freemium with high LLM cost, etc.)
- Generates Stripe-ready tier copy for your pricing page
Claude Opus 4.6 recommends
credit-basedHigh LLM cost variance + B2B SMB market → usage-based pricing aligns cost-to-serve with revenue. Suggested 3 tiers below.
Starter
$29
200 credits
- Email support
- 1 workspace
Pro
POP$79
800 credits
- Priority queue
- Analytics
- 3 workspaces
Business
$199
3,000 credits
- SSO
- Audit logs
- Unlimited
Why credits, not flat rate?
Your heavy users cost 4.8× more to serve than your light users. A flat $79 plan would give 41% margin overall but lose money on heavy users. Credits align price with cost.
See what one agent run actually costs
Describe your agent as an ordered list of steps. Pick a model per step. See the per-step breakdown — and which step dominates. Catch the expensive LLM call before it eats your margin.
- Per-step cost in mono — no hand math
- Live 2,000+ model catalog from OpenRouter + LiteLLM
- Margin verdict given the price you charge per run
Classify ticket
Claude Haiku
Retrieve KB context
GPT-4o-mini
Draft reply
Claude Sonnet 4.5
Quality check
GPT-4o-mini
Cost per run
$0.0276
4 steps · ~92% margin at $0.35
Breakdown
Step 3 dominates at 56%.
Try Claude Haiku here to cut cost 70%.
Who it's for
Built for the people who own the price
If you decide what an AI product charges, you need to know what it costs to serve. That's the whole job.
Indie AI founders
Shipping a wrapper or AI SaaS solo. Price it right before launch — not after the first scary model bill.
Agent builders
Multi-step agents with a model per step. Find the one call that quietly eats most of your margin.
Product & growth teams
Setting or revising the price of an AI feature. Replace guesswork with defensible unit economics.
7-day free trial · Founding pricing
Try every feature free for 7 days
Start your free trial today — no charge for 7 days. After the trial, you're locked in at the founding rate, the lowest price we'll ever charge. Cancel anytime during the trial and pay nothing.
AI pricing engine
Claude Opus analyses your cost structure + market and recommends flat-rate, usage-based, credit-based, or hybrid pricing — with specific tiers and reasoning.
Cost discovery
Map every cost: LLM tokens, infra, third-party APIs, labor. We score what you might be forgetting and build your real cost-per-user automatically.
Agent run simulator
Describe your agent as an ordered list of steps. See cost-per-run, per-step breakdown, and margin given the price you charge.
Scenario simulator
12-month margin projection. Adjust growth, churn, conversion, price. Spot break-even month and total profit before you commit.
Pricing page generator
AI-written tier copy + headlines, ready for Stripe. Generate the pricing page you ship the same day.
Live LLM pricing catalog
2,000+ models across OpenAI, Anthropic, Google, DeepSeek, Meta, Mistral and more — refreshed daily from OpenRouter + LiteLLM.
Free vs founding
| Capability | Free | Founding |
|---|---|---|
| Free margin check (/check) | Unlimited | Unlimited |
| Projects | — | Unlimited |
| AI pricing recommendation | — | Included |
| Cost structure engine | — | Included |
| Agent run simulator | — | Included |
| Scenario simulator | — | Included |
| Pricing page generator | — | Included |
| Founding price locked in for life | — | ✓ |
Skip the waitlist. Back the build.
Get instant access to the full pricing engine, simulator, and surveys. Founding price locked in for life.
No charge for 7 days · Card required · Cancel anytime · Secured by Stripe
FAQ
Questions, answered straight
Where do the cost numbers come from?
How accurate are the estimates?
My agent isn't a simple chatbot. Does it still work?
Why not just use a spreadsheet?
Is my cost and pricing data private?
Do I need a credit card to try it?
What happens after the 7-day trial?
Which model providers are covered?
// stop pricing on vibes
Find out what your AI really costs
Run the free 30-second check now, or start the 7-day trial and get the full toolkit — AI pricing engine, agent simulator, and scenario modeling.
no charge · 7 days · cancel anytime