Real AI agent use cases, with sources and setup prompts

Test an AI agent on a small investing account

Let an AI agent trade a small Robinhood account isolated from his other accounts.

$12.95/mo to start2-4 hoursBeginnerSource linked

Source

@EverydayAi10 on X · 2026-10-10

Written in our own words. Credit to the original author. Link checked 2026-10-14.

What they did

Let an AI agent trade a small Robinhood account isolated from his other accounts.

Result (as claimed)

$83 became $77 in about four days, so he paused it.

Tools they used

Grok BotRobinhood

Tools in our setup

Suggested in the setup steps below; not named in the source.

GitHubRailwayTradingViewAlpacaTelegram Bot APIClaude CodeMassive (formerly Polygon.io)

Cost: TradingView Essential $12.95/mo (billed annually) for webhook alerts. Alpaca paper trading, Massive Stocks Basic and the Railway trial cost $0 to start; Telegram Bot API pricing is not listed in its docs; coding agent from its own plan.

Set it up

  1. Create the repo and paper account

    Create a new Node.js (or Python) project in this repo for a TradingView-to-Alpaca webhook bot. Add a .env.example with APCA_API_KEY_ID, APCA_API_SECRET_KEY, APCA_API_BASE_URL=https://paper-api.alpaca.markets, WEBHOOK_SECRET, TELEGRAM_BOT_TOKEN, TELEGRAM_CHAT_ID. Never commit .env. Use the official Alpaca SDK and default to the paper endpoint; refuse to start if APCA_API_BASE_URL points to live unless LIVE_TRADING=true is set.

    Docs: docs.alpaca.markets/docs/paper-trading, docs.github.com/en/get-started/start-your-journe

  2. Build the webhook endpoint

    Add a POST /webhook endpoint that accepts TradingView alert JSON like {"secret":"...","symbol":"AAPL","side":"buy","qty":1}. Reject requests whose secret does not match WEBHOOK_SECRET, validate symbol/side/qty, respond within 3 seconds (TradingView cancels slower requests), and process the order asynchronously. Log every request.

    Docs: tradingview.com/support/solutions/43000529348-ab

  3. Place orders on Alpaca

    Implement order placement with the Alpaca Trading API (POST /v2/orders) using market or limit orders. Add risk limits from env vars: MAX_POSITION_USD, MAX_ORDERS_PER_DAY, and a kill switch TRADING_ENABLED. Check buying power and existing positions before each order. Write unit tests that mock the Alpaca client.

    Docs: docs.alpaca.markets/reference/postorder

  4. Backtest with market data

    Write a script that pulls daily bars for the strategy's symbols from the Massive API (MASSIVE_API_KEY) and replays the strategy rules to estimate returns and max drawdown. Print a summary table. Keep it simple and document the assumptions.

    Docs: massive.com/docs

  5. Send Telegram alerts

    Add a notify() helper that sends a message via the Telegram Bot API sendMessage method using TELEGRAM_BOT_TOKEN and TELEGRAM_CHAT_ID for every order, rejection and error.

    Docs: core.telegram.org/bots/api#sendmessage, core.telegram.org/bots/tutorial

  6. Deploy to Railway

    Prepare this service for Railway: add a start command, read PORT from env, and add a /health route. Then tell me exactly which env vars to set in the Railway service and how to copy the public HTTPS URL into the TradingView alert's Webhook URL field.

    Docs: docs.railway.com/quick-start

Safety: Not financial advice. Trading involves risk of loss. Paper trade first, for weeks, and review every trade. Keep API keys in env vars, use a webhook secret, and add position limits and a kill switch. TradingView webhooks require 2-factor authentication on your account.

Copy the full setup prompt

Paste into your agent and it will walk you through the setup, asking before anything that sends, posts or spends.

You are helping me set up this AI use case end to end in Claude Code.

Goal: Test an AI agent on a small investing account
What the original author did: Let an AI agent trade a small Robinhood account isolated from his other accounts.
Reported result (as claimed by the author): $83 became $77 in about four days, so he paused it.
Source: https://x.com/EverydayAi10/status/2108915470141173932
Tools they used: Grok Bot, Robinhood
Category: Money

Please:
1. Ask me 3 short questions about my situation (accounts I have, budget, what I want the output to look like).
2. Propose a step-by-step plan using these tools or cheaper alternatives, with the time and cost of each step. Use official pricing pages; say "check pricing" if unsure.
3. Then do the setup with me one step at a time, waiting for my OK before anything that sends messages, posts publicly, spends money or deletes data.
4. Finish with a checklist to run it again next time.

Related use cases