We just hit 100 followers
Thanks to everyone who’s been following the journey as we build agent-first payment infrastructure, live on Mantle mainnet
Your support means a lot. We are just getting started, more updates coming soon as we continue shipping.
Grateful to have you here ❤️
This one actually hits a bit too close for comfort.
I have been in that cycle building agent workflows, you hardcode a model because it’s decent at reasoning or tool use for one step, then a few days later something cheaper or better at a specific capability drops and you are back in the code swapping names, testing again, and hoping the rest of the flow doesn’t break, or you try to force one model to handle everything and it just underperforms on certain parts of the job. It gets old fast once you’re running real multi-step stuff.
Being able to write small policies instead, like “this step needs JSON mode and should be cheap” or “this one needs stronger reasoning and no logging” and have it route properly while still showing you exactly which models were considered, rejected, and why… that feels like a proper solution instead of another abstraction that hides the mess.
Especially when you want to actually understand and trust what’s happening across a workflow.
The fact that it’s open source and runs on your own keys makes it even more interesting. Nice ship.
▪️ @Qevorpay made it to the Top 30 finalists for the Turing Test Hackathon Demo Day
Out of 120 shortlisted projects, we will be presenting on July 3 in the Consumer DApp track.
Grateful for the journey so far. More updates coming.
🔥🔥
Qevor is building agent-first payment infrastructure on Mantle.
Made for both regular users and AI agents, with safety and usability in mind.
Any support on the community vote would mean a lot.
If this pops up on your feed quote with anything, this is our entry for the mantle turning social track
#MantleAIHackathon
Final submission for the @Bitget_AI Hackathon S1 🚀
This is VesperClaw, an accountable and self-improving AI trading agent built for Bitget.
VesperClaw scans BTC, ETH, and SOL, generates trade ideas, and puts every idea through multiple checks before allowing any paper trade. It refuses weak setups, protects capital during losing streaks, and actually learns from its closed trades.
The core idea is simple: a trading agent should not only show what it did, but also explain why it did it and why it refused certain trades.
Here’s what’s included:
➖ Full trading loop with perception, decision, execution, monitoring, and learning
➖ Alpha Gate and AgentVault as dual risk layers that check both individual trades and overall portfolio exposure
➖ Profit Guard that makes the agent more defensive after losses
Evolution memory that updates strategy behavior based on real closed trade results
➖ Meme Radar that analyzes trending meme coins and gives clear buy/watch/avoid signals
➖ Prediction market section for football and World Cup markets
➖ Full public paper trading logs and a live dashboard that shows every decision and refusal
Built using Bitget Agent Hub, Playbook, and Qwen.
Demo:
https://t.co/xCQEmqFDVe
Live dashboard:
http://38.49.209.149:8501/
#BitgetHackathon
Qevor is building multi-chain, agent-first payment infrastructure, with Mantle as our first mainnet.
It allows both everyday users and AI agents to create, manage, and execute payments safely.
Current features:
• Agent Workspace with AI copilot
• Safety checks & policy controls
• Agent Escrow with scoped balances
• Wallet history & stats
• Byreal preflight integration
#MantleAIHackathon
Qevor is building multi-chain, agent-first payment infrastructure, with Mantle as our first mainnet.
It allows both everyday users and AI agents to create, manage, and execute payments safely.
Current features:
• Agent Workspace with AI copilot
• Safety checks & policy controls
• Agent Escrow with scoped balances
• Wallet history & stats
• Byreal preflight integration
#MantleAIHackathon
I'm building VesperClaw, an autonomous Bitget paper-trading agent that doesn’t just execute trades, it explains every decision and learns from its outcomes.
Built for the @Bitget_AI Hackathon S1.
It turns multi-source market data into explainable Signal Mandates, runs them through a strict risk firewall, and evolves its strategy behavior after every closed trade based on what actually worked in each market regime.
More details and demo coming soon.
#BitgetHackathon
Qevie is officially live on @qieblockchain Mainnet.
An agent native QUSDC execution layer built on ERC-4337 smart accounts, scoped session keys, Qevie Paymaster, QIE Domain Resolver, AgentPolicyManager, ReceiptRegistry, and Passport.
Users can send, request, batch, create links, set recurring flows, or simply issue agent commands.
Autopilot maps intent to the right payment rail, checks policy constraints, resolves `name.qie` recipients, submits UserOps, and records execution receipts.
Policy defines the boundary:
- QUSDC-only execution
- allowed recipients
- per-tx / daily / weekly caps
- expiry
- gas behavior
- guardian revocation
Execution is automated only where the smart account allows it.
Try it: https://t.co/YpsxJrxZjY
OPEN TO FEEDBACKS AND SUGGESTIONS!
Here's what's live on Qevie now on @qieblockchain testnet.
🧵 👇
1/5
A QUSDC payment app on QIE with smart accounts, sponsored onboarding gas, wallet-based signing, and a mobile-first payment UX.
QIE Wallet signs. Qevie Smart Account executes
Introducing Qevor: The definitive USDC payment infrastructure for the @arc Testnet.
We are streamlining on-chain value transfer by merging scalable batch distributions and intuitive social payment links into a single, unified gateway.
A seamless standard for Web3 payments. 👇🧵
My core focus as an Architect is making the @arc ecosystem as frictionless as possible. Moving crypto shouldn't be stressful, so @jennyoliver57 and I built Qevor to fix exactly that.
To break down what we built simply:
🔹 Getting paid: Generate a custom link or QR code, share it anywhere, and get paid in native USDC.
🔹 Sending crypto: No more sweating over 42-character hex addresses, just use a simple Qevor username or .eth username.
🔹 Mass payouts: Need to pay a team or a DAO? Push funds to multiple wallets in one single transaction, safely verified by an AI co-pilot.
We set out to build the Venmo/Stripe of the Arc Testnet, and I'm super proud of what we shipped.
Dive into the thread below 👇
@samconnerone
The last 48 hours have been an absolute blur, but we finally made it to the finish line
Just shipped my official solo submission for the @XLayerOfficial hackathon. NGL, getting a local LLM to autonomously execute trades on mainnet without draining its own gas vault almost took me out, but it works.
I built @PrismPu7se because I wanted a real sovereign agent, not just a glorified chatbot wrapper. It runs completely on a private Ollama engine, scores signals from Onchain OS, and executes real micro-swaps to prove conviction. It even monetizes its own alpha via x402 paywalls to self-fund.
Attached a screenshot of the terminal running live, you can literally see it hit the low gas threshold, autonomously swap USDC to OKB to refill itself, and go right back to scanning for signals
Super proud of how this architecture turned out.
Will I be updating it and making it better? Yes ✅
Stay tuned
Check out the demo and the repo in the thread below
#BuildOnXlayer #BuildX
The intelligence layer of Web3 has arrived.
Introducing PrismPulse: The first private, fully autonomous asset manager and trading agent built natively on @XLayerOfficial
Welcome to the new meta. 🧵👇 #BuildX#BuildOnXlayer
Most people lose money in crypto for 3 reasons:
✅ Scams
✅ Bad research
✅ Fake communities.
I designed an AI system on @Mantle_Official that solves all three.
Introducing the Mantle AI Core Suite: Three interconnected, autonomous agents redefining L2 safety
Live demo:
https://t.co/v6dkV5JE8U
Here’s how it works 👇🧵