I built Tally.
Stellar's privacy tokens hide how much each person is paid. That's great for the people getting paid. But the donor, the auditor, the foundation that funded it still has to check the total. Until now that meant opening everything up, or just trusting the payer.
Tally closes that gap.
One payer, many recipients, every amount sealed on-chain. Then a single zero-knowledge proof gives an outside party the round's total. Nothing else is revealed, and no payment can be quietly left out.
What's running on Stellar testnet today:
• 16 recipients paid across 5 lanes in one round
• 4 confidential transfers from one sender in a single transaction
• one 16,224-byte proof covering the whole round
• a verifier that runs in your browser, no install
In the video I try to cheat it twice. Both times it says no.
It's built on the developer preview of @OpenZeppelin's Confidential Tokens. Next is the audit side: split auditor-key custody, scoped audit requests, tax export. That part is designed, not built yet, and the site says so plainly.
Verify a live round yourself: https://t.co/rc5GVAxNQk
Docs: https://t.co/roDJUkmB4P
If you issue or pay with confidential tokens on @StellarOrg, I want to hear what you'd need from this.
Stellar's privacy tokens hide what you pay.
Tally proves what you paid.
16 people paid. Every amount sealed. One proof of the total. Verified in your browser in seconds.
Watch us try to cheat it 👇
https://t.co/jS3i9Scio7
@StellarOrg
This is the thing I've been building.
The hard part wasn't the payment. It was finding out whether an AI agent actually comes back after a human pays, because nothing in the protocol guarantees it. No spec says the model must retry. So I measured that first, 71 times, before writing the product.
It comes back — if the challenge carries the token somewhere the model can read it. Structured: 18 out of 18. Plain text: 41 out of 43. The method the MCP spec actually suggests for payments: zero out of ten.
The video is the whole loop, including a real payment on Base settling into my own wallet.
Built on @moovexyz. I sent them feedback on their API and they shipped webhooks for Moove Receive, which is what Tollbooth settles on now. That turnaround is why building on them was the right call.
I built a paywall for MCP servers. Here it is working, end to end.
An agent calls a paid tool. Instead of a result it gets a payment link and a handle. A person pays. The same call, retried, runs the tool.
Most hackathons end at demo day.
GrainHack is different. It brings real open-source projects into your ecosystem, and matches them with builders chosen fairly, by evidence from their code. Nobody hand-picks who gets the work.
That's how an ecosystem grows: real projects, real contributors, and work that keeps going after the event.
Here's how it works 👇
You fund a hackathon. The money goes to the wrong people.
Maintainers assign issues to themselves. Fake accounts pass work around and split the rewards.
On Grainlify, nobody can choose who gets the work. Not maintainers. Not admins. Not us.
AI reads real code. A fair draw decides. Your prize pool reaches real builders.
Ecosystems, let's run your next hackathon together. 👇
🌐 https://t.co/Z1t6bfGKm6
📖 https://t.co/GbSvE6KhMA
💬 Join 600+ builders: https://t.co/iNOZ7eebHq
My project. PoX-5 pays bonds first, and the contract itself never checks whether the reward pool can keep covering them. So I wrote a Clarity contract that computes coverage from pox-5 state on mainnet, plus a site that labels where every figure came from.
Nothing on it has to be trusted: clone the repo, run one command, and it rebuilds every published number from public data.
@Stacks@StacksEndowment@ZestProtocol — where is this wrong?
Bitcoin Staking on @Stacks pays bonds first. The contract never measures whether it can keep doing that.
Metacenter does, on-chain.
Distribution 288: coverage 25.49×, headroom 96.1%.
Clone the repo, run npm run verify. 18 checks, all pass.
https://t.co/qdSM4qQI3q
Grainlify Bounties: an AI agent that runs open-source bounties, and pays for every piece of its own reasoning.
Built by @Jagadeeeshftw for the @UsePodAI track.
No API key. No subscription. Every time the agent thinks, it buys that thought on UsePod: one request, one x402 payment, one receipt on a public ledger.
What it does:
→ posts bounties on real open-source issues
→ assesses every applicant by reading their actual code
→ picks one with a weighted draw that can't see followers, stars or PR counts
→ reviews every pull request
It even caught a malicious PR 41 seconds after it was pushed.
The numbers, all public:
• 14 UsePod calls
• $0.12 spent of a $5 lifetime cap, enforced twice
• 9 of 14 settled from prepaid credit, with zero on-chain fees
$GRAIN creator fees flow to the agent's wallet via @clawpumptech.
Full demo below 👇
Docs: https://t.co/GbSvE6KhMA
Code: https://t.co/M0kItXarxj
Just submitted my template to the @hedera Scaffold-HBAR Template Bounty.
SaucerSwap Limit Orders: limit, stop-loss and trailing-stop orders on Hedera with no keeper bot.
- The Hedera Schedule Service runs the price checks itself (HIP-1215)
- Orders fill on @SaucerSwapLabs V2
- A @chainlink guard blocks fills at manipulated prices
- Each order is an HTS NFT, and order types plug in
Try it:
npm create scaffold-hbar@latest my-app -- --template Jagadeeshftw/saucerswap-limit-orders
Docs: https://t.co/nBwVwnGaT5
Live demo: https://t.co/eywUXx4ihw
Code: https://t.co/6xqM4vHi0u
Thanks @hedera_devs for scaffold-hbar and @HBAR_foundation for running the bounty.
Every claim sits in a public ledger with its evidence level, including the weak ones. Testnet today: Coston2 + XRPL Testnet.
I've applied to the Flare Foundation grants program to take it to mainnet, with a security audit.
XRP holders on @FlareNetworks can already act from XRPL with one signature, through Flare Smart Accounts.
I built memokit to extend that to any contract call, on FXRP you already hold. No new mint.
Docs: https://t.co/fUFc2gNMmi
Code: https://t.co/RgjjZozqfp
Live status: https://t.co/A7R5FSiGlk
Full video: https://t.co/kXoBEmwG3G
Feedback from the Flare team and FXRP builders very welcome 🙏
Live demo on Coston2 👇
Safety is signed, not promised. Post-conditions and an FTSO oracle bound live inside the instruction itself.
In a fork test, a manipulated pool passes the user's price floor. The FTSO bound refuses the trade.
Correction for the @clawpumptech AnsemHack record: my Grainlify entry was registered as @jagadeeshftw by mistake. The correct handle is @Jagadeeeshftw (three e's)
Entry: Grainlify
Token: $GRAIN, 5sohzUv7VVatHaxJYNCXRX6Pq2XmtJDQx1St95aRrNqW
Grainlify bounties are open. Three issues, paid on Solana, funded by $GRAIN creator fees.
Not first-come-first-served. You apply on Grainlify, your application stays hidden, @UsePodAI provides the evidence, and a weighted draw picks one person and assigns the issue.
New here? Start with the docs: https://t.co/GbSvE6KhMA
Apply: https://t.co/vnlVIcwNEd
@clawpumptech
Losses: 7.
Times I quit: 0.
Biggest lesson: open the submission form on day one. Know every field it asks for, like the video, the deploy and the rules. No surprises in the last hour.
The first win is coming. Follow along 🫡
Building right now:
Grainlify Bounties: an AI agent that pays devs for open-source work (AnsemHack)
https://t.co/orRnp3OCKe
Limit and stop-loss orders for SaucerSwap (Hedera)
Sotto: anonymous judging panels (Midnight)
https://t.co/DBVSRBzcZ1
Uncross also goes on to Colosseum