The story behind Tickerbot.

We’re David Spiro (co-founder & CEO) and Marc Auger (co-founder & CTO). We met as two of the first few employees at Bubble in 2018, helping build a visual programming language for shipping web apps without writing code. In 2022 we pointed no-code at algorithmic trading, a domain we’d been playing with on the side for fun. The side project turned into a company: we built nights and weekends around our day jobs until October 2025, when we both went full time. Today David owns GTM and product, and Marc owns DevOps and signal compute.

Every trading product we’ve shipped sat on the same stack: pull OHLCV from a feed, stand up a Postgres warehouse, compute the indicators and signals, index it all point-in-time so historical queries work, schedule the scans, wire up webhook delivery, then maintain the whole thing as data corrections come in. The feed was the easy step. The work was everything above it — and every build started the same way: weeks of plumbing before the first row came back.

The first was Tickerterm, a no-code platform for building and running trading algorithms — a hosted runtime that executed them for you, charts on par with TradingView, portfolio management, and, feeding it all, a warehouse of precomputed signals. When AI arrived, we pivoted the authoring to agent chat: explore ideas in conversation, generate the strategy code, hand it to the runtime. It worked, but it fought its own audience: people who build trading algorithms don’t want to run them in someone else’s stack. They already have execution through a brokerage, and they’re technical enough to manage their own runtime — they didn’t need ours.

So we took the technology underneath and put a mobile app on top to widen the audience: describe a setup in natural language, get a push notification when it fires. Same warehouse, same scans, simpler surface. It worked too, but we didn’t like who we were becoming: we’re technical, we wanted to serve a technical audience, and a consumer app is neither. We went back to the drawing board, looked at everything we’d built over the years, and decided to build what we wanted — not what we guessed others wanted — and what we wanted was what Tickerbot is today.

This is the third, and the right one. The warehouse, the SQL grammar, and the webhook delivery had done the real work inside both prior products; this time they are the product. Every US equity indexed, the signals precomputed, SQL on top. That’s Tickerbot — the stock market, in SQL. And we think it’s working, because it turns out there are a lot of people like us — people who just want to write the query, not rebuild the stack underneath it.

Rows are tickers. Columns are signals. The warehouse covers ~14,602 tickers — every US equity, plus the US Treasury curve, major FX pairs, spot metals, and major crypto pairs. Each row carries ~269 precomputed fields (price, volume, RSI, MAs, ATR, gap math, position metrics, volatility) and ~152 condition flags (gap_up, at_52w_high, breakout, rsi_oversold, …) — everything a market participant asks about, already a column. Each signal refreshes on its own cadence: live ones every minute during market hours, fundamentals daily, financials when issuers file.

Compose any mix of signals into a SQL WHERE clause — one grammar, wherever a WHERE clause applies — and every read runs live, as of any past moment, or as a webhook subscription. Ten endpoint families front it:

• /v2/tickers — one symbol’s full state, plus ETF holdings and sectors
• /v2/signals — every ticker matching a named signal
• /v2/scan — any SQL query across the whole market
• /v2/series — ticker state history over time, plus OHLCV bars
• /v2/events — everything that happened, on one timeline: dividends, splits, insider transactions, analyst actions, news, signal firings
• /v2/news — the news archive, filterable by sentiment, recency, or source
• /v2/signals (write) — author custom signal columns in the same SQL grammar
• /v2/universes — named ticker lists, custom or system-provided
• /v2/webhooks — subscription registry; signed, retried, idempotent delivery
• /v2/stream — live updates over a held-open websocket

Under the hood:

• Scanner — Node + PostgreSQL on DigitalOcean
• API gateway — Node/Express on DigitalOcean; auth and billing on Firebase
• Webhook scheduler — fires the moment each data refresh lands (real-time) on every paid plan; optionally throttle a webhook to hourly or at market open
• History retention — as-of queries on every plan, Free included.
• Indicator validation — every formula checked against its published definition, guarded by a suite of automated regression tests
• Delivery — signed (HMAC SHA-256), retried on 5xx, idempotency keys on every event