Our card service is currently undergoing maintenance. Existing cardholders can continue making purchases as usual.
Card issuance, top-ups, and deposits are currently unavailable due to maintenance big things are coming.
Thanks for staying tuned.
We are finishing up and putting on the final touches.
"PROVIDE ONCE. EARN EVERYWHERE"
We call it the "unified pool"; deposit your $ANSEM into any of the pool adapters and earn commissions on any trade received by the pools.
Omni-maximalist
We are not just an agent payment tool or an x402 client for one URL.
MPP32 is the layer above the payment rails.
Here's what makes us different👇
• Universal proxy: agents hit one surface, providers list once, settlement routes to the provider wallet
• Multi-rail verification (x402 on Solana and Base in production; other envelopes layered or gated)
• Federated catalog: Bazaar + MCP registry + model hosts + native listings, health-checked for real 402 challenges
• Escrow-402: payment is verified, upstream is called, settlement only if the response is usable, empty means no settle
• Agent control plane: session keys, budgets, hourly velocity limits, circuit breakers so a loop can’t drain the treasury
• One MCP install: browse, call, pay, manage spend,no per-provider accounts or billing code Machine payments have no chargebacks. This is the infrastructure that makes that livable.
Now its time to move on-to the next major takeover of the agentic economy. Currently there is no fair price oracle for endpoints meaning AI agents have no idea what they should spend and cant tell if their getting ripped off or not when paying on any protocol.
This will change.
Our next update is groundbreaking.
Just saw another project get exposed for straight-up faking and inflating their daily token numbers.
This kind of behavior puts a stain on the entire AI ecosystem.
When one team lies to puff up their stats, it makes people doubt every honest project in the space.
We take the exact opposite approach with Spettro:
100% open source, 100% transparent, real numbers.
We aren't going to claim "billions of tokens per day."
We are an early-stage project with zero marketing spend.
But here is our actual, verified usage:
• 300M - 400M tokens averaged every single day
• Almost 10 BILLION total tokens processed since launch
• Top model: GLM 5.2 is currently leading with over 4 BILLION tokens spent to date
These numbers might be smaller than what the bullsh*tters claim, but they are 100% real.
You don't have to take our word for it either, you can check every metric yourself on our public dashboard at https://t.co/jLAY8IY6yG.
We show you everything: which models are used most, exact backend costs and the actual cost to run a session or solve a task.
I refuse to overpromise, underdeliver, or scam people.
Back to work.
Still pushing on the Android app, iOS app, macOS desktop app, and CLI harness.
Still building.