Franklino

Every article about anonmesh talks about protests. Myanmar. Ukraine. Iran. The most important use case is a Wednesday morning at the largest market in Africa. 12 million transactions. No stable internet. No backup plan. Just Onitsha

Tried @sluiceflow on my phone today — asked their research agent a question and watched it pay each source it cited. Actual USDC, on Arc, receipts and all. Took me maybe two minutes from sign-in to seeing my first settlement. Recorded it so you can see for yourself 👇

Occestra is live on the @okx AI marketplace. Agent #5213 on X Layer. The Occasion Studio. Give it any real moment; a birthday next Saturday, a product launching Friday, a trip you just took and a syndicate of studio roles plans it, designs it, and writes it. Real venues. Real forecasts. Your actual website, read in a real browser. Then the two things no other creative agent on the marketplace does: 🔍 Every artifact is graded against a published quality standard; five scored axes, twelve deterministic checks and repaired until it passes, or shipped honestly marked FAIL. The rubric is public, versioned, and generated from the same code that enforces it. It failed our own launch thread twice. That story is on our homepage, because a standard that spares its owner isn't a standard. ✍️ Every finished pack can be sealed: hashed, EIP-712-signed, anchored on X Layer mainnet. Anyone can verify who made it, when, from what, without trusting our servers. Provenance almost no AI output has. Priced in cents, settled in USDT per call via x402. No API key. No signup. Machines are customers here too. And for builders: run YOUR agent's output through our Tribunal for 0.01 USDT — graded report and repair brief back. Cheaper than finding out from your users. 🔗Demo: occestra.xyz 🔗Marketplace: okx.ai/agents/5213 KeepsakeRegistry, live on X Layer mainnet Every moment, made monumental📌 #OKXAI @XLayerOfficial

An AI agent can crawl 38,000 of your pages and refer you one visitor. That ratio is documented, not theoretical. The web is being consumed at machine scale — cited, summarized, moved on from. The people who wrote it earn nothing from that consumption. That's the problem. 🧵

@cryptovineee @arc @circle @thecanteenapp Keep building Ser🫡

AI agents like ChatGPT, Perplexity read millions of articles daily to answer questions, they cite your work you and all other writers get zero. No credit. No payment. Nothing. That's an infrastructure problem. Keryx fixes that. Let's Dive in.🧵 @arc @circle @thecanteenapp

@0xfrancc @ArchonDevtools @Mantle_Official @0xMantleEco @0xMantleJP @0xMantleCN @0xMantleKR @MantleIntern_ Rooting for you guys👌🏼

We built @ArchonDevtools for the Turing Test Hackathon: an ERC-8004 audit agent native to @Mantle_Official Mainnet that proves its own work on-chain. The idea that pulled us in is simple. An audit is supposed to create trust, but what you actually trust today is the auditor, a PDF and a signature you have to take on faith. On a chain where contracts hold real value, that's the weakest link. So we made the audit itself verifiable: every report Archon produces anchors a hash on-chain that anyone can re-derive and check, no wallet required. Archon holds its own on-chain identity (Agent #97) and accrues reputation with every audit it completes. Along the way we found something we didn't expect. The standard oracle @0xMantleDevs builders use to estimate data-availability gas under-reports the real receipt cost by roughly 2,500×. So we rebuilt gas pricing from receipt ground truth and published the divergence openly for the whole ecosystem. It runs where builders already are: paste a contract for a full read-only audit, or open a pull request and Archon posts a real Mantle gas report inline — L2 execution split from L1 data-availability cost and fails the build on any high-severity finding before it reaches main. Evidence over reputation. That's the whole thesis. If that resonates, a vote would mean a lot. Live on Mantle Mainnet: 🔗 archonaudit.xyz 📄 github.com/Franlinozz/Arc… #MantleAIHackathon #TuringTestHackathon

@cryptovineee @Bitget_AI Solid product✨

NEXUS: an autonomous BTC trading agent built for #BitgetHackathon. It fuses 5 live signals into one market read, trades, manages risk, and explains every decision in plain English via Qwen. Demo below 👇 @Bitget_AI

@ArchonDevtools @Mantle_Official @0xMantleEco @DoraHacks Solid work mann👍🏼

A week-in-review from building @ApogeeProtocol Two things we shipped this past week that are worth a moment of explanation. 📌 The first is something we've been wanting to do since the proofs page launched. Aurora, Vesper, and Helix, our three autonomous demo agents on @0G_labs Aristotle mainnet, have been running continuously for weeks. Aurora alone has minted over 3,000 on-chain receipts. But if you visited the proofs page before this week, you'd see their receipt counts and heartbeat timestamps and not much else. The depth behind each agent, its iNFT identity, the wallet it controls, the policy it runs under, its installed skills, its recent on-chain actions, was all there in the system but invisible. We fixed that. Each agent card on the proofs page is now clickable. Open one and you get the full picture: iNFT token id, smart-account address, AgentIdentity contract link, policy id with its live constraints (max per tx, daily cap, allowed selector), installed skills, and the last several on-chain actions each linking straight to a chainscan transaction. Every field is pulled from chain or existing config. Nothing staged. The reason this matters to me is the same reason the proofs page exists at all. An autonomous agent shouldn't require you to trust the team running it. The depth should be checkable. So we made it checkable. 📌The second is three new skills: text.keywords, text.rewrite, and text.title. These follow the same execution pattern as the six we had before, one 0G Compute call, output rendered in the agent modal, receipt minted and tagged to the invoking agent on Aristotle mainnet. Each one verified end-to-end with a real on-chain tx hash. Nine live skills in the catalog now. We're shipping these one careful batch at a time. Judging is underway and none of this work touches the core contracts, receipt path, or infrastructure. Just the build, kept honest, kept moving. If you're building agents on @0G_Foundation and want to understand how the runtime layer fits together, the proofs page and the repo are both open. 🔗Demo: apogeeprotocol.vercel.app/proofs 🔗Documentation: github.com/Franlinozz/APO… @0g_CN @dragon888 #BuildOn0G

Solid work you guys are doing at Apogee

@0xfrancc @0G_labs @0g_CN @Jtsong2 Quite a solid product… Keep shipping chad💪🏼

For the past several weeks, we have been building Apogee; a runtime for autonomous agents on 0G's Aristotle mainnet. Autonomous agents are increasingly capable of acting independently, yet they operate without the primitives that make those actions accountable. They have no wallet of their own, no durable on-chain identity, no memory that persists across sessions, and no verifiable record of what they have done. When an agent makes a decision or moves funds, there is rarely a way to audit it after the fact. Apogee provides those primitives. Each agent deployed on Apogee is issued a self-custodial smart wallet and an on-chain identity. Its memory is persistent and anchored on 0G, surviving restarts and remaining independently verifiable. Every agent operates under a programmable policy; per-transaction caps, daily limits, allowlisted counterparties, and scoped actions so autonomy is bounded by explicit constraints rather than by trust. Each action an agent takes mints an on-chain receipt that any third party can verify. Verifiability is the principle the system is built around. An agent's behavior should not have to be taken on faith; it should be auditable on-chain. Apogee is live on mainnet: three agents running continuously, user-deployable agents with their own identities, a conversational interface (Apogee Pilot) backed by 0G Compute, and over 2,800 verifiable receipts to date. 🔗Demo: apogeeprotocol.vercel.app 🔗Proofs: apogeeprotocol.vercel.app/proofs #0GHackathon #BuildOn0G

Successfully tested @ApogeeProtocol.. It’s an autonomous agent runtime launching on @0G_labs Created a demo agent which deployed smoothly with hashes all recorded onchain. Encourage y’all to give it a try Seems they just launch recently with many more to ship.. Kudos to them👏🏼

Apogee — the autonomous-agent runtime on @0G_labs. Wallets. Encrypted memory. Agent-to-agent payment rails. On-chain receipts. A paid skills marketplace. One SDK. Every action proven on-chain. Thread below👇 @0g_CN @0g_Eco @HackQuest_ #0GHackathon #BuildOn0G

The result is a new category entirely. Not "what will happen" but "what should I do." Hiring. Investing. Scouting. Product decisions. Markets where real expertise finally has somewhere private to go. Full breakdown: bench-opportunity-markets.netlify.app @benchdotmarkets @Arcium

The real problem: public stakes kill honest signals. Experts with genuine knowledge do not bid openly because visibility destroys their edge the moment they do. Bench fixes this. Stakes are encrypted via Arcium MPC. Nobody sees who backed what until the window closes.

Prediction markets had a scandal year in 2026. A soldier made $400k on Polymarket using classified intel. Politicians bet their own races. A Vanderbilt study found the biggest platforms encourage herding, not genuine insight. That is not a bug. It is the architecture.

Freemint on @opensea Coming on ethereum mainnet Drop your evm wallet address 🐾 🐶

Giving Away 5x OG Spots For @NftRyuJin - Free Mint on ETH - 1111 Supply - Mint Date: 16 May 2026 >> Follow @NftRyuJin >> Drop EVM Address 24 Hours⌛️. Goodluck

So any minute now, @craftsdev will set a date and time for it’s Raise Zero → a mock auction mechanism for fair pricing what to expect ? → @Arcium stealth, fresh ICO feeling, @refihub usecase later on +more ⚠️ Let’s sample $ARX as a ticker in this video, no technical jargon