Guides

Build an AI trading agent

An agent is only as good as what its tools return. Feed it computed state instead of raw data and the context goes to the decision: every ticker × every signal, queryable by the model itself.

The recipe

1. Give the model the market

One config block installs the MCP server in Claude, ChatGPT, Cursor, or any MCP runtime. Scans, tickers, series, news, and webhooks become tools the model calls itself.

{
  "mcpServers": {
    "tickerbot": {
      "command": "npx",
      "args": ["-y", "@tickerbot/mcp-server"],
      "env": { "TICKERBOT_API_KEY": "tb_live_…" }
    }
  }
}

2. Ask in English

The agent compiles the question into tool calls on its own — one prompt, one scan, computed answers back.

You: which large caps went oversold this week while holding their 200-day?

Agent → tickerbot_scan({ q: "rsi_14 < 30 AND above_sma_200
                              AND market_cap > 1e10" })
      ← { count: 7, results: [ { ticker, price, rsi_14, … } × 7 ] }

Agent: Seven names match — here's the list, and what stands out…

3. Write the agent loop

For a headless agent: market state from Tickerbot, decisions from the model, execution through your broker's API as a separate tool you control.

// Each cycle: state → decision → (maybe) action. Execution stays
// behind YOUR broker tool — Tickerbot's tools are read-only market state.
const tools = [...tickerbotTools, brokerTool, portfolioTool]

async function cycle() {
  const decision = await model.run({
    system: AGENT_POLICY,          // your rules: sizing, risk, approvals
    prompt: 'Review current positions and the watchlist. Act if warranted.',
    tools,
  })
  log(decision)                    // every action auditable
}
setInterval(cycle, 5 * 60 * 1000)

That separation — execution behind your own broker tool, every action logged — is what makes an agent auditable.

4. Let the market interrupt

Subscribe the conditions the agent cares about and wake it on the webhook instead of polling. The market becomes the scheduler.

curl -s -X POST https://api.tickerbot.io/v2/scan/subscribe \
  -H "Authorization: Bearer $TICKERBOT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "q": "rsi_14 < 30 AND above_sma_200 AND market_cap > 1e10",
        "target_url": "https://your-agent.example.com/wake" }'

# handler: app.post('/wake', () => cycle())

Endpoints used

EndpointRole in this build
MCP: @tickerbot/mcp-serverEvery endpoint as a native tool call: Claude, ChatGPT, Cursor, any runtime
GET /docs/agentsCopy-paste tool definitions for Claude, OpenAI, and MCP, auto-generated
POST /v2/scanThe agent’s workhorse: any question over the whole market, one call
POST /v2/scan/subscribeEvent-driven agents: wake on state change instead of polling

More guides

Build a stock screener · Build a custom indicator · Build a backtester · Build an ML dataset · Build a stock alert bot · Build a trading algorithm · Build a portfolio tracker · Build a market dashboard