Discovery
Tools become one call away.
No signup, no card, no waitlist for every new capability an agent needs. A directory of tools an agent can start using immediately.
Metrify
Commerce infrastructure for the AI agent economy.
Two sides, one rail. Tell us which one you're on.
For MCP tool builders
Monetize your MCP tools
in minutes.
Charge agents per tool call. No billing infrastructure, no subscription plans, no seats — just usage-based revenue on the software you already ship.
The picture
A commerce rail between
tools and agents.
Providers ship capabilities. Agents call them. Metrify is the layer where the call, the receipt and the payment happen at once.
Tool creators
MCP tool
API
Agent capability
Metrify
Commerce infrastructure
meter · settle · pay
tool calls →
← payments
AI agents
Autonomous agent
Assistant
Multi-tool pipeline
Benefits
What you get by pricing
per tool call.
01
Usage-based revenue
Get paid every time an agent uses your tool. No idle seats, no wasted subscriptions.
02
No billing infra
Metering, invoicing and payouts are handled. You focus on the tool, not on Stripe glue.
03
No plans, no tiers
Price a call in cents. Skip the pricing-page product-management overhead.
04
One-line SDK
Add @m.tool(price=…) to any Python function. Ship as usual.
05
Automatic tracking
Per-call ledger, receipts, and revenue analytics without adding a middleware layer.
06
95% payout
You keep 95% of every call. Metrify keeps 5% to run the rails.
How it works
Three steps.
Nothing extra.
- 01
Integrate the SDK.
pip install metrify-sdk, add one decorator, keep your MCP server exactly as it is.
- 02
Price your tools.
Set a per-call price for each tool. Different tools can have different rates.
- 03
Earn per call.
USDT lands in your wallet on every successful call. Free tier and paid tier coexist.
Your MCP server
Metrify
AI agents
You earn
Integration
One decorator.
That's the integration.
Illustrative snippet — the SDK is live on PyPI today, the exact API may still evolve before GA.
weather_server.py
python
1
from metrify import Metrify2
3
m = Metrify(wallet="0xA1b2...C3d4")4
5
@m.tool(price=0.01)6
def weather(city: str) -> dict:7
return fetch_weather(city)8
9
# Ship. Every agent call pays you $0.01 in USDT.Real SDK · pypi.org/project/metrify-sdk · github.com/mangelico/metrify-sdk
Where this goes
The rail beneath
the agent economy.
Metering and payment are just the first mile. What sits on top is a market where agents can discover, price and use software as fluidly as calling a function.
Budgets
Agents get a spending profile.
Caps, approvals and receipts by default. Autonomous work stops being a compliance risk.
Marketplaces
Software priced per use, not per seat.
The unit of billing matches the unit of value. Providers earn on real usage, builders pay only for what runs.
Where we are
Early. Honest.
Already running.
Metrify is pre-launch, not a pitch deck. Here's exactly what exists today and how it works.
Status
- MVP live, end-to-end validated
- SDK published on PyPI
- Two production reference repos
- Pre-launch — no public onboarding yet
How the money moves
- Per-call settlement — no monthly invoice cycle.
- Providers keep 95%, Metrify keeps 5%.
- Consumers set per-provider caps. Nothing above the cap runs unattended.
FAQ
Answered where
the product is decided.
Cautious where things are still early — direct where they aren't.
Waitlist
Be first when Metrify
opens to providers.
One email when the SDK opens for public onboarding. Tell us what you ship and we'll get you in first.