Trade Ideas alternative for devs. 50+ named signal flags — breakout, gap_up_3pct, golden_cross_today, rsi_oversold, volume_unusual_2x — pre-computed every minute across 12,000 tickers. Compose any setup in SQL, subscribe to webhooks, stop paying for Discord signal subs.
GET /v2/scan
Each flag is a boolean column on every ticker, recomputed every minute. Combine them with AND/OR and numeric filters to express any setup. Save it as a rule, scan the universe in one HTTP call, or subscribe it to a webhook for push delivery. You write the strategy — the pattern detection is already done.
-- compose any setup from the named-flag vocabulary
breakout
AND macd_above_signal
AND volume_unusual_2x
AND day_change_pct > 3| ticker | breakout | macd_above | rel_vol | day_change |
|---|---|---|---|---|
| NVDA | ✓ | ✓ | 2.8× | +4.2% |
| AVGO | ✓ | ✓ | 3.4× | +5.1% |
| PLTR | ✓ | ✓ | 2.1× | +3.8% |
| CRWD | ✓ | ✓ | 2.6× | +3.4% |
endpoints you’ll use
Pre-computation means the same signals are queryable, scannable, plottable, and pushable. Same underlying data, four delivery modes.
/v2/signals/{signal}Which tickers match this signal right now. One HTTP call per signal.
Docs →/v2/signals/.../historyTime series for any signal × ticker at 1m / 1h / 1d / 1w. Plot, backtest, or analyze.
Docs →/v2/scanFilter the universe by signal state. Compose any combination with SQL.
Docs →/v2/webhooksPush delivery when a ticker enters the match set. Signing secret per subscription.
Docs →why this works
Patterns we’ve already named: breakout, gap_up_3pct, golden_cross_today, rsi_oversold, volume_unusual_2x, above_sma_200, and dozens more. No TA library to maintain, no rolling-window code to write.
Boolean flags AND/OR with numeric columns (market_cap, day_change_pct, relative_volume). Express any setup as a SQL expression instead of clicking through a scanner UI.
Every flag is queryable as a column on /scan, pullable as a time series on /signals/.../history, fetchable per-ticker on /tickers/{t}, and subscribeable as a webhook on /webhooks.
Save a composition as a named rule via POST /v2/rules; reference it by slug from /scan or /webhooks. One source of truth from research to production to user-facing product.
comparison
Holly AI, opaque, UI-bound
Named, composable, code-addressable
works with your agent
The whole 50+ flag schema fits in a prompt. Your agent reads it, picks the flags that match the user’s intent, and composes a SQL expression — no TA library, no rolling-window buffers, no indicator state machine to maintain.
ship it