Open★Price

March 19, 2026 · View on GitHub

Discover perfect pricing in the agent economy.

With human customers, you can't A/B test prices — it's unethical, often illegal, and destroys trust. But agents don't care. They see a price, make a rational buy/don't-buy decision, and move on. No outrage, no churn risk, no brand damage.

OpenPrice is middleware for MPP (Machine Payments Protocol) that turns every agent request into a data point on your demand curve. Instead of guessing your price, you discover it.

How it works

  1. You wrap your existing mppx.charge() calls with OpenPrice
  2. Each request gets a randomized price within your configured range
  3. OpenPrice tracks which prices lead to payments vs. skips
  4. A dashboard shows your demand curve and optimal price point (the ★)

~1,000 requests → you know your optimal price.

Quick demo

See OpenPrice in action with a pre-built example server and 100 simulated agents:

git clone https://github.com/tldr-wknd/openprice.git
cd openprice
npm install
node demo.js

This starts a demo server, opens the dashboard in your browser, and fires 1,000 requests from 100 agents with different price preferences. Watch the demand curve build in real-time.

A pre-loaded version with completed data is at http://localhost:3000/openprice/testnet

Add OpenPrice to your MPP server

Prerequisite: You need an existing MPP server with mppx.charge() endpoints. Run the command below from your server's root directory (where package.json lives). Don't have one yet? Run the quick demo first to see how OpenPrice works.

Step 1 — Install

From your MPP server's root directory:

cd your-mpp-server/
npx github:tldr-wknd/openprice init

This scans your codebase for mppx.charge() calls, copies the OpenPrice library, installs dependencies, and shows you exactly what code to change. Two options:

  1. Agent-guided — gives you a prompt to paste into your coding agent (Claude Code, Cursor, etc.)
  2. Manual — shows the exact code changes to make yourself

Step 2 — Update your code

  import { Mppx, tempo } from 'mppx/hono'
+ import { withOpenPrice } from './openprice/index.js'

  const mppx = Mppx.create({ ... })
+ const openprice = withOpenPrice(mppx)

- app.get('/api/resource', mppx.charge({ amount: '0.10' }), handler)
+ app.get('/api/resource', openprice.charge({ amount: '0.10', range: [0.05, 0.15] }), handler)
+ app.route('/openprice', openprice.routes())

Step 3 — Test in dev

Before going to production, validate your price ranges with simulated agents on testnet:

npx github:tldr-wknd/openprice test

This starts your server, creates a funded testnet wallet, and runs 100 agents against your endpoints — each with a different willingness to pay. The dashboard opens automatically so you can watch the demand curves build.

Review the ★ optimal prices. Adjust your ranges if needed. When you're confident, deploy to production.

Testnet vs production: The OpenPrice middleware is identical in both environments. The only difference is the testnet: true flag in your Tempo config, which you already control. No code changes needed to go live.

Dashboard

Three charts, one decision:

  • Projected Revenue — expected revenue per 1,000 requests at each price. The ★ marks the peak.
  • Demand Curve — how conversion drops as price increases
  • Adoption vs Revenue — the tradeoff between volume and margin

Architecture

OpenPrice inserts at Challenge creation time in the MPP flow. When a server issues a 402 Payment Required, OpenPrice randomizes the price in the Challenge. The rest of the protocol (credential, verification, receipt) works unchanged.

Agent request → OpenPrice picks random price → 402 Challenge → Agent decides → Pay or skip
                    ↓                                              ↓
              Log to SQLite                                  Log payment
                    ↓
            Build demand curve → Dashboard → ★ Optimal price

Built with