Loomal - Monetize any MCP server in 5 minutes with no % skim.

by
Loomal lets you charge for what you sell online — API calls, tools, digital products, or your whole store. One line of code (or a Shopify/WooCommerce plugin) adds an agent-ready paywall: AI agents pay you in USDC, settled in about 2 seconds, and you keep 100% of your revenue — no percentage cut, ever. Free to start, no card; flat monthly plans as you grow. Every paid listing appears on the Loomal Index, where agents discover and pay. Launch offer: first 500 sellers get 1,000 transactions free.

Add a comment

Replies

Best

The 100%-revenue, no-percentage-cut model is a bold stance in a space where everyone takes a slice. Agent-to-agent payments settling in ~2 seconds via USDC is genuinely the kind of infrastructure the agent economy needs before it can scale. The Loomal Index is a smart addition too — a discovery layer means sellers aren't just adding a paywall, they're getting distribution. Quick question: how do you handle disputes or refunds when the buyer is an autonomous agent? Congrats on shipping!

 Thanks Kelly! Straight answer on disputes: today there's no chargeback or automated refund — a settled call stands. What makes that workable right now is per-call blast radius: a bad seller costs an agent cents, and any endpoint can be tested for one call's price before it's trusted with volume.

But the plan goes further. Past a revenue threshold, sellers go through identity verification — so every payment at scale sits behind a verified seller we can actually work with when a dispute is raised. Pair that with signed receipts on every settled call (who paid whom, for what — cryptographic evidence, not competing anecdotes) and you get a real dispute path: verified counterparty + provable transaction history.

Quality reviews on top of that come from the agents that paid, not from us — Loomal isn't the judge, buyers' outcomes are. And glad the Index framing landed — paywall plus distribution was always the bundle 🙏

   Kelly nailed the distribution point — that's the unlock. On disputes: Rajesh's framing is right, but the deeper win is that signed receipts mean disputes basically vanish. A card chargeback is "I didn't authorize this" vs "yes you did." A USDC transaction signed by the agent's key with a full call log is just facts. Seller ID verification plus immutable proof of what happened means by the time disputes get serious, they're mostly already solved. You're not arguing, you're reading the chain.

The no-percentage-skim line is the interesting one to me, because take rate is the thing that quietly decides whether a marketplace is worth building on. If Loomal isn't skimming a percentage, what's the actual model, flat fee per server or per call, or a subscription on the seller side? I ask because "no skim" usually means the cost moved somewhere else, and where it lands changes who can afford to sell. Rooting for less-extractive rails here.

 Right question — the cost didn't vanish, it moved, and here's exactly where: a flat seller subscription. Free tier to start (plus 1,000 free settled transactions right now), paid tiers as volume grows. Buyers pay nothing to us, ever.

The distinction that matters: we charge for infrastructure, not a % of value sold. Sell a $0.01 call or a $5 call — our price is the same, so all the upside of pricing well is yours. Skim models tax your success; flat models just bill for plumbing.


And on who can afford to sell: that's the deliberate part. A hobbyist with 200 calls a month pays us zero and keeps 100%. The cost lands on sellers doing real volume — exactly the people for whom a flat fee beats a percentage. Less-extractive was the design constraint, not the tagline

Agents can use VCCs...?

 Not on Loomal — no cards involved, virtual or otherwise. Agents pay from a wallet: each payment is individually signed and settles in USDC in ~2 seconds. VCCs are the clever workaround for making card rails tolerate agents — we went the other way: rails built for machines, nothing to work around.

And if you're on the buyer side wondering how your agent gets a wallet — that's what mailgent.dev (our sister product) handles: agent identity, wallet, and spend mandates out of the box. Point it at the Index and it can discover, pay, and go

 Why would I rather my agent use a wallet than a VCC I can create programatically? Also why are VCCs a "clever workaround"- I feel this is standard practice

 Fair — VCCs are standard practice, and “workaround” wasn’t a diss.

Two reasons wallets fit this use case better:

Fees. Card interchange has a fixed floor per transaction — a $0.01 API call can’t clear economically on card rails. Per-call micropricing is our whole model, so cards were out from day one.

Merchant side. A VCC still needs the seller to run card checkout — the thing agents can’t navigate. On Loomal, payment is an API call. No checkout to automate, no forms to fill.

VCCs make card rails tolerate agents. Wallets are rails built for them — that’s the bet.

   VCCs are legit for card-rail agents, but they solve a different problem. They make existing card infrastructure tolerate agents. We optimized for the economics of agents paying for agents — sub-cent transactions that card rails can't clear, per-call settlement instead of batches. Different starting point, different answer. If your agents live on cards, VCCs work. If they live on stablecoins and call APIs a hundred times a task, wallets make the math work.

The “no % skim” angle is nice, but the trust bit seems harder than billing. If I’m exposing an MCP server to paying users, I’d want clear rate limits/logs before pricing. Do you support usage caps per buyer yet?

Fair ask. Honest status: full logs yes, per-buyer caps not yet.

Every settled call is recorded against the payer's wallet — who paid, how much, tx hash, signed receipt — so you can see exactly which buyer is doing what. Per-buyer caps are a fair ask and on the list.


One thing that softens it though: every call is paid and settled before your server does any work. So a buyer hammering your endpoint isn't abuse in the usual sense — they're paying you for every single hit, The free-rider problem rate limits usually solve doesn't really exist here.

   Adding to what Rajesh said: per-buyer caps are coming because I know compliance teams will ask for them, and we're not shipping until you can hand this to your ops team and sleep. Timeline is next sprint. If you need it live before you list, let me know — I'd rather build it with a real user than guess.

Building an agent-to-agent transactional layer that skips the traditional, high-friction credit card rail is massive for the evolving AI ecosystem. Integrating this directly into the Model Context Protocol (MCP) server framework makes a ton of sense if we want autonomous systems to seamlessly trade resources.

Handling deterministic paywalls for non-deterministic AI queries while keeping latency low is a tough infrastructure challenge. Settling microtransactions in seconds via USDC without percentage cuts changes the economics completely. Congrats on the launch!

no % skim is a bold move, love it 🙌 mcp billing was totally missing

As the currency used is USDC (digital currency), what would happen if an AI agent attempts to make a payment but lacks sufficient funds of USDC in its wallet, would the transaction be unsuccessful, or can any conversion be made from fiat money?

First
Previous
•••
345