Build Story · Crypto futures screener · Experiment · EN
Signal Hub: a live crypto futures screener
A free screener that scores buyer pressure across crypto futures and sends alerts to a dashboard and Telegram.
In one look
The problem
In crypto futures, buyer pressure builds fast and quietly. By the time you have clicked through a watchlist, the move is over.
What we built
Signal Hub reads public market data from Binance futures, DeFiLlama and CoinGecko. It rolls buyer-side volume, a volatility score and a flow shift into one composite score per pair, across four timeframes, and raises an alert when a score crosses. No login and no exchange keys.
Where it is now
Live and free, with no login. It is an experiment, built in Lovable in short bursts between March and October 2026, with a server-side backtest of its signals and an export for your own analysis. The last edit, on 3 October, added a daily bounce summary. Not financial advice.
What this would have cost in 2024
About $7.9k and 5–8 weeks
2024 team: 2 people: a full-stack developer and a part-time front-end designer. Built here by one founder working with AI in 17 working days over 7 months.
| Piece | 2024 hours |
|---|---|
| Futures dashboard: multi-timeframe table, charts, DeFi TVL radar, data health | ~74 |
| Composite score, scheduled scans, Telegram alerts | ~57 |
| Backtest engine and signal export | ~51 |
| Landing page, brand pass, share card | ~29 |
| Total (PERT) | ~210 |
Method: PERT (O+4M+P)/6 per piece; low = O × $25, likely = PERT × $37.5, high = P × $50 (Accelerance 2024 Eastern-Europe band). These are estimates of what a team would have needed in 2024, not invoices.
How it was built
From first commit to current build in 17 working days over 7 months
- The dashboard, in one night
A futures table with volume changes over several timeframes from Binance, a TVL radar fed by DeFiLlama and market-cap dominance charts, then a full UI overhaul.
32 edits in about 4.5 hours - Scheduled scans and Telegram alerts
Scans run on a timer and a Telegram bot posts alerts into a group, each with a timeframe badge and price. A composite-score explainer and tooltips on every field went in the same day.
24 edits in about 5 hours - Backtests, an export and a security pass
A backtest function was added on 5 July and an export on 6 July. On 23 August a results engine arrived, which tests the signal formulas against historical prices on the server. A security pass followed five days later.
- A landing page and a brand
A landing page took the front door and the dashboard moved to its own page. It has a lock-on hero, a scan animation, a logo and a share card.
6 edits in one day
Who did what
The founder decided
- Public market data only: no login and no exchange keys.
- Magenta is for fills and buttons. Green is kept for data, and falling candles stay red.
AI built and checked
- The app was built in Lovable from the founder's prompts: 82 of the 83 edits came from its AI agent.
- The landing page, hero art and share card came from a design session with Claude, starting from a generated brand poster.
Stack
Want something like this for your team?
From one call to something live. The same method is in the Claude Operator Pack.
From the founder
You might also like
Shares React · Edge functions
Shares React · Edge functions
Shares React · Edge functions