Gradience

AI-native payment, trading and settlement infrastructure with deterministic workflow safety.

OpenClaw-native strategy tools Small-cap live safety gates LI.FI quote + status tracking

Quick Install Watch Demo Flow Ops Runbook

Install in 60 Seconds

Install extension directly from GitHub source.
Reload Pi.
Start with natural-language read/compose/execute intents.

pi install https://github.com/DaviRain-Su/pi-chain-tools

Sui local wallet is loaded from ~/.sui/sui_config/sui.keystore and active address from client.yaml.

What Gradience Is

Agent-first toolset covering multi-chain read / compose / execute / rpc groups.
Workflow model: analysis → simulate → execute, with risk-aware prompts and deterministic outputs.
Cross-chain execution surface includes SOL/SPL, swaps, staking, Polymarket BTC5m, NEAR DeFi/Intents, and CETUS-based Sui basics.
Designed for AI automation: natural language intent parsing, guarded mainnet execution, structured summaries, and machine-readable validation artifacts.

Stablecoin Yield AI Agent

Maximise your stablecoin yield with your own AI Agent that analyses and allocates into DeFi protocols on autopilot 🤖💸

通过您专属的 AI 代理，在 🤖 💸 上自动分析并分配到 DeFi 协议中，最大化您的稳定币收益。

示例意图可直接在 Pi 中试跑：

“帮我分析我当前的 USDC、USDT 和 DAI 持仓，并给出风险分级后的稳定币收益分配方案。”
“把我 20% 的 USDC 转到收益更高、风险较低的借贷池，剩余保留高流动性备用金。”
“对比当前 DeFi 利率，模拟迁移到更高 APR 的协议并输出安全警示。”

Monad AI Hackathon (Track 1)

Gradience is positioned as Track 1：原生智能体支付与基础设施 — a foundational layer for AI-callable chain settlement, payment, and transaction capabilities.

Provides structured service discovery through toolset/intent registry.
Supports one-step quality checks for agent deployment via scripts.
Supports billing/settlement integration via execution artifacts and runId-phase events.

Competition dossier and English/Chinese versions: monad-ai-hackathon.md / monad-ai-hackathon.en.md

Latest Updates (2026-02)

ACP Async Queue Durability: queue snapshots persist across restart and auto-resume queued jobs safely.
Payment Lifecycle APIs: /api/payments/create, /api/payments/confirm, /api/payments.
Execution Hard Gate: strategy execution now requires buyer + paymentId + paid status + strategy/buyer match + active entitlement.
Replay Protection: duplicate terminal runs with same runId + paymentId are blocked.
Ops UI Upgrade: ACP filters/detail panels and a new Payments panel for pending/paid/failed visibility.

Read details: near-dashboard.md · core-architecture-principles.md · openclaw-acp-quickstart.md · openclaw-ops-runbook.md

Production Readiness Checklist

✅ ACP async queue persistence + restart resume
✅ Strategy execution hard gate: buyer + paymentId + paid validation + entitlement
✅ Replay protection: duplicate terminal run blocked by runId + paymentId
✅ ACP/Payments operational visibility in dashboard
✅ Async retry backoff + max-attempt + dead-letter queue (DLQ)
✅ BSC execute path supports native RPC signer mode + command fallback
🟡 Next: real payment provider callback integration hardening (provider-specific signing/idempotency schema)

API Examples (ACP / Payments)

Use these examples with dashboard server running on 127.0.0.1:8787.

curl -s -X POST http://127.0.0.1:8787/api/payments/create -H 'content-type: application/json' -d '{"confirm":true,"strategyId":"stb-usdc-v1","buyer":"demo-buyer"}'

curl -s -X POST http://127.0.0.1:8787/api/payments/confirm -H 'content-type: application/json' -d '{"confirm":true,"paymentId":"pay-REPLACE","txRef":"tx-demo"}'

curl -s -X POST http://127.0.0.1:8787/api/acp/job/execute -H 'content-type: application/json' -d '{"confirm":true,"dryRun":false,"runId":"acp-demo-1","strategyId":"stb-usdc-v1","buyer":"demo-buyer","paymentId":"pay-REPLACE","requirements":{"targetChain":"near","intentType":"rebalance","amountRaw":"1000000"}}'

AI/CI Quality Loop

For reliable agent usage, all schema and runtime contracts are validated through CI-gated scripts:

npm run schema:check-files — strict file manifest (human)
npm run schema:check-files:json — strict manifest (JSON)
npm run schema:ci-check — one-step manifest + full check
npm run schema:audit — strict JSON diagnostics for automation and AI workflows

npm run schema:audit

Natural Language Examples

Solana / Sui / NEAR / EVM

帮我查一下 Sui 主网本地钱包余额（包含USDC）

Swap Simulation

把 0.01 SUI 换成 USDC，先模拟。

Polymarket BTC5m

买 BTC 5分钟涨 20 USDC，保守，先模拟

Mainnet Execute

继续执行刚才这笔，确认主网执行。

Demo One-Click Commands

Demo 流程（点击即可跳转并高亮对应区域）：

演示前建议先执行以下命令（可一键复制到终端）：

npm run lint

npm run schema:audit

npm run check

npm test

npm run security:check

Show install from GitHub source.
Run npm run schema:audit before launch.
Run one cross-chain intent in analysis/simulate/execute flow.
Show structured execution summary + mainnet safety checks.