Web2’s N8N made automation feel like Lego; Web3’s been missing that glue. The Vision drop from @Talus_Labs finally clicks it into place: a visual, node-based layer on top of Nexus so you can drag → connect → deploy straight to Sui basically “N8N for web3,” per their post announcing Vision’s editor and direct deploy flow. Under the hood it’s still the serious stack: Nexus formalizes Tools and composes them into DAG workflows, with Leaders coordinating off-chain steps and settling results on Sui/Move (resource model = real asset authority). The whitepaper lays out why this matters for agents that need to hold keys, pay for calls, and leave a verifiable trail. ❯ Try the flow: sketch a mini pipeline, wire it in Vision, then ship to Sui. ❯ Sanity-check that Leader events and fee draws match expectations; tweak tool pricing if usage spikes. ❯ If you want code receipts, the nexus-sdk shows active work, which pairs neatly with a visual-first builder surface. --- My read: this reduces the “custom glue” tax that kills most on-chain automation. If Vision ships stable deploys and keeps typed IO strict, you get fewer brittle bots and more composable agents. I’ll be watching deploy latency and how fast teams move from a canvas prototype to a paid workflow if those curves tighten, Vision isn’t just a demo tool; it’s the on-ramp to production AvA agents.