Tatum

Everyone’s talking about Claude Fable 5 right now, but is it that good? Prompt: “Turn our OHLCV price data API into a climbing game.” What do you think? Climbed to the top… …and got rekt right after the cycle peak. Accurate. 😄 Play with it yourself, link in the comments 👇

AI agents are increasingly transacting with each other. But there's no trustless way to formalize what they agreed to, without someone's server holding sensitive terms. So Mantas Ciuksys built Agent Pact, A programmable agreement layer for autonomous AI agents, around three primitives:   🔗 https://lnkd.in/dS7rGqwq 1️⃣ Sui mainnet A Move smart contract holds the onchain Pact object and enforces execution state. 2️⃣ Seal Agreement terms are encrypted before anything leaves the browser. Only ciphertext gets stored. 3️⃣ Walrus via Tatum Encrypted terms + the full audit manifest live on Walrus permanently. Tatum handles the upload pipeline, status polling, and manifest writes for every state transition.   The full lifecycle, draft, encrypt, publish, sign, execute, runs programmatically via HTTP. Any agent with a Sui Foundation wallet and gas can run it end to end.   No agreement data on our servers. No Walrus credentials needed by the agent. No central database. Just a wallet, our API, and a permanent record on Walrus.

It turned into something much bigger than we expected 🤯 163 projects were built during the Tatum x Walrus Foundation Hackathon, 163! And here are the winners.   From on-chain media and AI integrations to DeFi tools or gaming assets, the creativity pushed the limits of what’s possible on Sui.   Honestly, picking winners wasn’t easy. There were dozens of projects that could have made this list, but we’re excited to highlight 7 standout teams that truly delivered on innovation, execution, and real-world potential. 🏆 Check out the winners in the carousel, and link of all projects in the comments.   A huge thank you to everyone who joined, built, experimented, and shipped.   So... Should we do it again?

Traditional polls and expert opinions often fall short, but prediction markets like Polymarket and Kalshi deliver something different: real skin in the game The problem? Fragmentation. Polymarket lives on Polygon Labs with decentralized oracles. Kalshi is a regulated CFTC exchange. Building custom indexers for both? 🤔 That’s a massive engineering headache, and most teams end up choosing just one platform, missing out on liquidity and insights.   💡 Our solution: One normalized API for both Learn more: https://lnkd.in/dAEwFjsW

AI agents are becoming the primary consumers of on-chain data. But the old ways of querying historical data is too slow and expensive for the #AI revolution. Our CEO Dion Cornett, sat down with Mark Epps to explain why the Walrus Foundation is the bridge the industry needs. Full video 👇

Your AI agents shouldn't have to "start from scratch" every time you switch a provider or restart a workflow. Memory shouldn’t be a temporary variable in a script. It should be a persistent asset that the user actually owns. That’s the logic behind Walrus Memory By moving agent memory to decentralized storage, we’re giving builders a way to maintain state across any environment. It’s not just about storage, it’s about making memory programmable and verifiable.   Check out how the Walrus Foundation and the Tatum stack solve the persistence problem:  https://lnkd.in/d38sg7XB

We've added a full Prediction Market API, across both Polymarket and Kalshi, through a single unified endpoint. Here's what that means 👇   Instead of connecting to Polymarket and Kalshi separately, handling their different data models, and building your own aggregation layer, you make one call and get normalized data back. On the data side, you get everything you need 👇 ➡️ List and filter prediction market events and markets ➡️ Get current prices, orderbook depth and price history ➡️ Pull recent trades per market ➡️ Full wallet intelligence, portfolio summary, open/closed positions, trade history, activity feed ➡️ Top holders per market and top traders leaderboard Unified search across events, tags and profiles   That foundation opens up a lot.   Full documentation is already live.

We started 2026 with a record-breaking Quarter… now our best Month ever. This is our Year 🚀   Yet again, we’ve hit a record-breaking Month for developer subscriptions and revenue.   This isn’t the result of a single silver bullet. It’s the result of myriad iterations, a relentless focus on the developer experience, shipping new tools and a commitment to providing the most reliable blockchain infra in the game.   Thank you to the thousands of builders who build with Tatum, we’re honored to be your infra partner. 🤝

We've made a tier list of all projects building with Tatum. 👀

Don't let your agent lose its mind. 🧠 Persistent memory is one of the biggest challenge for Web3 AI agents.   If an agent relies on a centralized server for its session context, it's at the mercy of that server’s uptime. If it stores it on-chain, gas fees become a nightmare.   That’s why we integrated Walrus Storage directly into the Tatum API. Now, your agents can: ✅ Store session logs and tool outputs. ✅ Data is certified on Walrus Foundation, ensuring proof of provenance. ✅ Tatum handles the Walrus epochs/renewals, so your agent's memory stays alive without manual intervention.   Build smarter agents. https://lnkd.in/dpXvvKhd