← All posts
Product· 3 min read

TickerPilot: An AI Copilot for Stock Research in Telegram

A Telegram bot that turns a ticker into a full research pass: live news, analyst sentiment, and a technical read with an enforced risk:reward, then lets you interrogate the report instead of just reading it.

I’ve been heads down on TickerPilot for two weeks. It’s a Telegram bot for US stocks: type a ticker, get back a research pass that would normally take opening five tabs and reading for ten minutes.

The problem it’s solving

Doing real due diligence on a ticker before you trade it means pulling from three different places every time: news (what’s actually moving the stock right now), analyst sentiment (are targets getting raised or cut), and a technical read (where’s the actual entry, and what’s the real risk on this trade). Most people either skip most of that and trade on a headline, or do it properly and burn twenty minutes per ticker they’re just casually checking.

The other failure mode is the risk side. It’s easy to eyeball a stop and target and end up with a risk:reward that doesn’t actually make sense, because nothing forces the math to check out. TickerPilot enforces it: the stop and target it returns have to clear a minimum risk:reward, or it doesn’t hand you the number.

Why a single prompt wasn’t enough

The first version was one model call: stuff in the price data and news, ask for a report. It worked, but the quality wobbled depending on what the model decided to focus on that run. Tickers with a lot of news would get thin technical sections, and vice versa.

The current version runs through several agents instead of one. Separate agents handle price action, news context, and risk sizing, and their outputs get stitched into one report at the end. Splitting the work this way made the output noticeably more consistent, each agent has one job instead of juggling all three.

The part I use the most: /ask

After a report comes back, /ask lets you follow up on it. Ask “why this stop and not tighter” and it actually answers based on the specifics of that ticker’s report, instead of restating the summary. This is the feature that made the bot feel useful to me personally rather than just a report generator.

Rounding it out

  • /analyze: full report on a ticker
  • /ask: follow up questions on the last report
  • /watchlist: track tickers without burning API quota on every check
  • /signals: scan the watchlist for daily setups
  • /alert: get pinged when a price level hits

Watchlist and alerts exist so you’re not manually rechecking the same names every morning. The bot watches; you get pinged when something’s actually worth looking at.

Where it’s at

Still solo, still shipping fast, still figuring out what people actually want versus what I assume they want. Code’s on GitHub: daily_stock_analysis and ticker-pilot-bot. If you trade US equities, search Telegram for TickerPilotBot and try it.