tickerbot.io
View as markdown
Build guides

Custom signal

Tickerbot is the stock market API designed for agents. A custom signal is a named SQL expression on it: define it once, then use it like a built-in, ask for its history, and subscribe to it.

How it works

Chart-platform indicators (Pine Script, ThinkScript) are invisible to everything else you build. A custom signal lives in the API: define it once as a SQL predicate and the name works everywhere the built-in signals do.

1. Define it

A boolean SQL predicate over the 452+ built-in signals, or over other custom signals you own, so signals compose.

curl -X POST "https://api.tickerbot.io/v2/signals" \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "name": "oversold_quality",
        "expr": "rsi_14 < 30 AND above_sma_200 AND pe_ratio < 20",
        "description": "Oversold pullback in a profitable uptrend" }'
Show responseHide response
Response
{
  "as_of": "2026-08-21T11:01:52Z",
  "signal": {
    "name": "oversold_quality",
    "kind": "custom",
    "description": "Oversold pullback in a profitable uptrend",
    "expr": "rsi_14 < 30 AND above_sma_200 AND pe_ratio < 20",
    …
  }
}

2. Use it like a built-in

Screen the whole market for it.

curl -X POST "https://api.tickerbot.io/v2/scan" \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "q": "oversold_quality AND market_cap > 1e9" }'

3. Ask for its history

Custom signals mix into /v2/series; transitions_only returns just the rows where it flipped: its full on/off history, as if it had always existed.

curl "https://api.tickerbot.io/v2/series?tickers=DVA,GTX&columns=close,oversold_quality&interval=1d&from=2024-01-01&transitions_only=true" \
  -H "Authorization: Bearer YOUR_KEY"

DVA entered on 2026-08-05, before this signal was defined. As-of reads work the same way: define it today, query its value at any past date.

4. Wire it to a push

Because it's a signal, it subscribes like one.

curl -X POST "https://api.tickerbot.io/v2/scan/subscribe" \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "q": "oversold_quality AND market_cap > 1e9",
        "target_url": "https://your-app.example.com/hook" }'

The subscription stores the expanded predicate. The custom name compiles into its SQL at create time.

Show responseHide response
Response
{
  "id": "wh_q-s2lOSjLTM",
  "q": "(rsi_14 < 30 AND above_sma_200 AND pe_ratio < 20) AND market_cap > 1e9",
  "channel": "webhook",
  "target_url": "https://your-app.example.com/hook",
  "signing_secret": "whsec_…",
  "status": "active",
  …
}

The calls behind it

CallWhat it does here
POST /v2/signalsCreate the signal: name + SQL predicate; composes with other customs
POST /v2/scanScreen on it across the whole market, mixed with any built-in
GET /v2/series?transitions_only=trueIts flip history: every on/off edge on an aligned grid
POST /v2/scan/subscribeThe signal as a push: fires when a ticker starts matching