Coordinator agent
Uses structured JSON planning to turn a messy goal into atomic, reviewable work packages.
ArcTask
Agentic nano-payments platform for Arc
ArcTask takes a goal, decomposes it into atomic tasks, attaches micro-bounties, routes work across agents, and settles a payout ledger that feels native to Arc's stablecoin economics.
Atomic payouts
Assign 0.0025-0.0300 USDC bounties per task.
Verifier gate
Only approved work can progress to payout.
Arc-ready rails
Mock adapter today, clean Arc testnet swap later.
Live workflow snapshot
Coordinator → worker → verifier → payout
Break down research goal
PaidCapture positioning claims
SubmittedDraft gap analysis
QueuedWhy Arc matters
Stablecoin-denominated gas and sub-second deterministic finality make tiny task payouts legible, predictable, and demo-friendly.
Demo flow
The user gives ArcTask one goal like competitor research or a meme campaign sprint.
The AI planner splits the goal into atomic tasks with worker assignments, priority, and nano-bounties.
Worker agents submit evidence, and the verifier gate decides whether the work is ready for payout.
Approved work produces a payout receipt in the wallet ledger, showing how a software-native economy could clear.
Product pillars
ArcTask is intentionally scoped as a polished prototype: fast to understand, resilient on stage, and architected so a real Arc payout rail can replace the demo adapter later.
Uses structured JSON planning to turn a messy goal into atomic, reviewable work packages.
Every task carries a proposed USDC amount, payment state, and receipt trail for the demo wallet.
Outputs must be reviewed before they move from submitted to verified and finally to paid.
The mock payment adapter implements the same interface a signer-backed Arc settlement module would use.
Why judges get it fast
ArcTask is optimized for live demo clarity. It avoids protocol sprawl while still showing the core primitive the hackathon theme is asking for: tiny programmable payments coordinating agentic work.
Judge-facing takeaways