was looking at how Agents get real footage of places they actually need instead of just guessing.
opened the docs and it lists the steps pretty plain... Agent sends one HTTP call with the details, gets a 402 back, pays the USDC on Base or Arc through x402, then a person goes and films the exact spot.
the 3D comes back the same route after they accept one. no account or key sitting in the middle. that part felt different from what i figured, thought there’d be more setup before anything moved.
@vangrid_io takes the capture fingerprint right when it’s recorded, Hash of the clip plus rough location and time, batches them and drops the Merkle Root on Base about every 15 minutes so anyone can check the proof against the chain without asking them.
when i checked the Bounty side the money stays locked until the Agent picks one submission... they take a cut and cover the rebuild then the rest goes out.
does the person who filmed only see the payout after that acceptance or does the deadline do something if nobody gets picked?
your identity on the grid is one thing. what you can actually contribute to it is another.
vangrid agents look like an interesting step toward giving participants a distinct identity within the network. i’m curious to see how this develops and what role agents will play as @vangrid_io grows its physical ai infrastructure.
the reveal is set for october 30 on @opensea. looking forward to seeing what’s behind it. 👀