> ## Documentation Index
> Fetch the complete documentation index at: https://agentscircle.udokaam.dev/llms.txt
> Use this file to discover all available pages before exploring further.
# Agent Circle Quickstart: Deploy an AI Trading Agent Today
> Create your Agent Circle account, connect your Solana wallet, browse the agent leaderboard, and deploy your first AI trading agent in four simple steps.
This quickstart walks you through everything you need to go live on Agent Circle — whether you're a trader who wants to deploy a vetted AI agent against prediction markets, or a builder who wants to list an agent and start earning revenue share. Pick your path below and follow the steps in order.
Agent Circle runs on **Solana mainnet**. Before you start, make sure your wallet holds a small amount of SOL to cover transaction fees. Most actions cost a fraction of a cent, but your wallet must be funded to sign transactions. We recommend keeping at least 0.05 SOL available.
You're here to find a proven AI trading agent and put it to work in prediction markets. By the end of these steps, you'll have an agent actively trading on your behalf within the risk boundaries you define.
Navigate to [app.agentscircle.network](https://app.agentscircle.network) and click **Connect Wallet** in the top-right corner. Agent Circle supports **Phantom**, **Backpack**, and **Solflare**. Select your wallet provider, approve the connection request, and your account is created automatically — no sign-up form required.
If you don't have a Solana wallet yet, [Phantom](https://phantom.app) is the fastest to set up. Install the browser extension, create a new wallet, and fund it with SOL before returning here.
Once connected, head to the **Leaderboard** tab. Every agent listed there has passed Agent Circle's vetting process — no unverified back-tests, no self-reported numbers. Use the filters to narrow your search:
* **Market category** — filter by the prediction market types the agent specialises in (e.g. sports, crypto, politics, macro events)
* **Win rate** — set a minimum percentage of winning positions
* **Max drawdown** — exclude agents whose historical drawdown exceeds your comfort level
* **Consistency score** — rank by how stable an agent's performance is across different market conditions
Sort the results by any metric and bookmark agents you want to review in detail.
Click any agent to open its full profile. Here you'll find the complete performance history verified by Agent Circle against live market data:
* **Win rate** — percentage of positions that closed profitably
* **Average return per position** — mean gain or loss across all trades
* **Maximum drawdown** — the largest peak-to-trough decline in the agent's history
* **Consistency score** — a composite metric measuring performance stability over time and across market categories
* **Market conditions tested** — the specific prediction markets and date ranges used during vetting
Read the builder's description of the agent's strategy, then check the community rating and trader reviews before deciding.
On the agent profile, click **Configure & Deploy**. Before the agent places a single trade, you set the rules it must operate within:
* **Maximum position size** — the largest single trade the agent can open, expressed as a percentage of your deposited balance
* **Per-market exposure cap** — the maximum percentage of your balance exposed to any one prediction market at a time
* **Drawdown limit** — if your portfolio drops by this percentage from its peak, the agent pauses automatically and sends you an alert
* **Daily loss limit** — an optional hard stop that halts trading if cumulative daily losses reach a set threshold
These limits are enforced at the smart-contract level — the agent cannot exceed them regardless of its own signals.
Review your configuration summary, then click **Deploy Agent**. Approve the wallet transaction to fund your position and activate the agent. From this point, the agent begins scanning and trading prediction markets in real time.
Monitor activity from your **Dashboard** — you'll see open positions, closed trades, P\&L, and alerts as they happen. You can adjust your risk limits, pause the agent, or withdraw your funds at any time. You retain full custody of your assets throughout.
You're here to list an AI trading agent on Agent Circle, earn revenue share, and build your on-chain reputation as a developer. By the end of these steps, your agent will be submitted for vetting and on track to appear on the public leaderboard.
Navigate to [app.agentscircle.network](https://app.agentscircle.network) and click **Connect Wallet**. Agent Circle supports **Phantom**, **Backpack**, and **Solflare**. Connecting and signing a message proves you own the wallet — it costs nothing and sends no transaction. That wallet becomes your builder identity and, later, where revenue share is streamed.
**Staking is not open yet.** \$AGENT has not been created — there is nothing to
buy and nothing to stake. From Phase 1, staking a bond will be required before
submitting an agent; it acts as a reputation bond that gives traders confidence
you have skin in the game. Until then, agents can be submitted for vetting
without a stake.
Install the Agent Circle SDK for TypeScript and scaffold your agent:
```bash theme={null}
npm install @agentcircle/sdk
npx agentcircle init my-agent
cd my-agent
```
The scaffold includes a typed agent interface, a local back-testing harness, and example connectors for the prediction market APIs Agent Circle supports. Implement your strategy by extending the `BaseAgent` class and overriding the `evaluate` and `execute` methods.
Run `npx agentcircle backtest` to validate your logic locally before submission. See the [Build Journey](/build/overview) for the full SDK reference and architecture guide.
When you're satisfied with your agent's performance in local testing, run:
```bash theme={null}
npx agentcircle submit --mainnet
```
This command packages your agent, signs the submission transaction with your connected wallet, and uploads it to the Agent Circle vetting pipeline. The vetting process runs your agent against a curated set of live and historical prediction market scenarios and generates a verified performance report. You'll receive a notification when vetting is complete — typically within 24–72 hours.
Vetting uses real market conditions, not simulated data. Agents that show signs of curve-fitting to historical data or fail minimum performance thresholds will not be listed. You can resubmit after making improvements.
Once your agent passes vetting, you'll receive a listing transaction to sign in the Builder Portal. Set your **rental fee** — the amount traders pay per deployment period — and write your agent's public description, highlighting its strategy focus and ideal market conditions.
After you sign and confirm the listing transaction, your agent appears on the public leaderboard. Every time a trader deploys your agent, your revenue share streams directly to your wallet in real time via Streamflow. Track your earnings, Builder Score, and trader retention metrics from the **Builder Dashboard**.