Tilt after losses
Your win rate falls while your size climbs after a losing streak. The broker shows neither.
Self-hosted trading journal for crypto futures
A journal you
actually own.
It rebuilds your Binance USDⓈ-M futures fills into real round-trip trades and shows you why you really lose.
- Self-hosted, runs anywhere
- Read-only broker keys
- Your P&L never leaves your box
Binance shows you what filled. Not that you lose to yourself.
Your broker gives you a flat list of fills. It never tells you that your win rate collapses after two losses, that one coin carries your whole month, or that you keep trading past the line you swore you'd stop at.
Hidden concentration
One ticker, one trade. Drop it and the edge is gone, but the equity curve hides it.
Give-back
You were green and let it run flat. The fills don't measure the profit you handed back.
Broken discipline
Past your daily stop, over your size cap, in your no-trade hours. Nothing flags it.
NextScalp
The radar
Looks forward and outward, at the market, before and during the trade. It finds the setup.
NextScalp Journal
The mirror
Looks back and inward, at you, after the trade. It shows what you did with it.
The radar finds the trade. The mirror shows what you did with it.
From raw fills to an honest read.
One pipeline, end to end, all of it on your own server. Your secrets never reach a browser.
-
01
Your fills
Read-only Binance USDⓈ-M futures fills, income and funding.
-
02
Round-trips
Scale-ins and partial exits collapse into one real trade per cycle.
-
03
24h context
Each trade tagged with volume, prior move and its MFE / MAE.
-
04
Your Postgres
Stored in your own database. There is no vendor backend.
-
05
The dashboard
Server-rendered analytics, read straight from your data.
The mirror, up close
Free journals draw your equity curve. This shows your habits.
The behavioral read no flat ledger can give you, scored from your own trades and your own journal tags.
Revenge, tilt and the trade after a loss
Back-to-back re-entries into the same symbol where size grew after a loss. The trade you take after a losing streak, with its average net, size and win rate. When size climbs while net falls, that is tilt, quantified.
- Revenge episodes, counted and drillable
- Tilt after N consecutive losses
- Losses: planned vs blown
Give-back, FOMO and R
How much of the favorable move you actually took before you exited. Net and win rate by how far price had already run before your entry. Results in R against your planned stop, with the profit you left on the table.
- Capture efficiency and "gave it back"
- FOMO / chasing buckets
- R-multiples and left-on-the-table R
Exit Lab
Most journals show what you made. This shows what you left on the table.
Every exit, replayed on your own trades, never on a model's imagination. See the profit you gave back at the exit, and which exit rule would have kept it.
The give-back tax, and your own plan
The total you were up at each trade's peak and did not keep, shown next to your real capture. Then: if every trade had simply exited at the target and stop you wrote in the journal, what would the book have made? Your plan net, beside your actual net, so you see whether your discretion helped or hurt.
- Give-back tax in dollars, with an honest ceiling
- Trade-your-own-plan net vs your actual net
- Where give-back is worst, by setup and symbol
Sweep every exit, replay any trade
A grid scores every fixed target-and-stop over your trades and highlights the best, in R or in percent. On any single trade, drag the target, stop, trailing stop, break-even and time-stop and watch the outcome recompute live on its real candles. A sandbox: nothing is saved, nothing is placed.
- Fixed-policy sweep, scored in R or percent
- Capture curve: where holding for more stops paying
- Live per-trade exit replay, all on your own machine
The dashboard
Every surface, one honest read.
Grouped into Analytics, Control and Journal, with a shared time range, 7 days to all-time, that windows every metric on the page.
Overview
Net after fees, win rate, profit factor, equity curve, drawdown, win-rate by hour and weekday.
Edge
Performance by setup, tail concentration, friction, P&L by volume, hold time and symbol.
Behavior
Revenge, tilt, capture, FOMO, P&L by emotion, planned vs blown losses, R distribution.
Exit Lab
Give-back tax, trade-your-own-plan, a fixed-policy sweep, the capture curve, and live exit replay.
Discipline
Your own rules, scored by frequency, not outcome. Strict mode flags the trades you never journaled.
Journal
Everything you tag by hand: coverage, P&L by setup, entry and exit reason, emotion, and an expectancy heatmap.
Private AI, your key
Two AI lenses that never touch a vendor's server.
The only feature that ever leaves your box, opt-in and through your own Anthropic key. Leave the key unset and it simply stays off.
Per-trade audit
One tap reads the trade's own numbers, your journal entry and your discipline rules, then explains thesis versus execution. It streams in, renders as markdown, and is saved per trade.
- Grounded in MFE / MAE, fees, funding and your tags
- Replies in your language, saved for the next visit
- Re-analyze or delete anytime, nothing automatic
A private coach, with memory
Lift the lens to the week or month: a review that names your recurring patterns, grades how well you kept last period's rules, and sets a short, concrete list for the next one, all grounded in that period's real numbers.
- Weekly and monthly, aligned to your timezone
- Rules for next period, saved and graded next time
- The whole review history stays in your database
Yours, end to end.
Not read-only on someone's cloud. Custodian-free: there is no someone. Install it on a server, a VPS, or your own machine, and run it for as long as you like.
Your data, your database
Trades, P&L and journal notes live in your own Postgres. No vendor backend, nothing to leak.
Read-only keys
The sync only needs read access. It can never place, modify or withdraw. Ever.
Secrets stay server-side
Keys and the database client never reach the browser. A bundle scan proves it on every build.
Admin-only access
A Telegram-id allow-list gates every route. No identity provider, no auth tables.
Docker out of the box
App and Postgres bundled, migrations applied on start. One command and your env.
Binance, read-only
Built for USDⓈ-M perpetual futures: funding, 24h context, liquidity, 24/7 sessions.
Honest by design
It is built to tell you the truth, including the parts that sting.
- Net, not gross. Headline numbers are after fees and funding, the money that actually hit the account.
- Thin samples are de-emphasized. A one-trade hour cannot masquerade as an edge.
- Concentration is surfaced. The math shows when your edge is really one trade or one coin.
- Discipline is outcome-blind. A profitable rule-break is luck, not vindication.
- Give-back is measured. Capture and left-on-the-table R quantify the profit you didn't keep.
Per trade
The whole trade, in one place.
Price action, your fills, your intent and the outcome, side by side on every round-trip.
Interactive candlestick
OHLC over the trade window with every fill marked, average entry and exit, MFE / MAE, and your planned stop and target when set.
Assisted journaling
One-tap tag suggestions from data you already own: trend vs counter-trend entries, exit-at-plan, a spontaneous re-entry after a loss. Nothing is auto-applied.
Merge the re-entries
Scale-ins and add-backs fold into one logical trade, counted once everywhere. The dashboard even spots the candidates, reversible with one split.
Make it yours.
Your language
Ships in English and Russian. Add your own in two guided steps, no code.
Your rules
Set every discipline threshold, toggle each rule, and turn on strict mode.
Your look
Strict monochrome by default, with an opt-in trader-green skin and dark themes.
Your timezone
Pick any timezone and every timestamp, the win-rate heatmap, the P&L calendar and daily streaks recompute in it.
See yourself in the demo.
A read-only instance loaded with a realistic few months of trades, every page populated. Walk the whole thing end to end before you ever point it at your account.
Explore the live demoDemo data only. No account, no keys, nothing to sign.
Early access
Not on sale yet. Get in line.
It is in active build toward a self-hosted release on a subscription. Want a seat when it opens, or have a question? Send a note.
Email journal@nextscalp.comSubject line: Early Access request.
Straight answers.
Do you ever see my data?
No. It is self-hosted. Your trades, P&L and notes live in your own database. There is no vendor backend to send them to.
Which exchange does it support?
Binance USDⓈ-M perpetual futures, for now. It is built deep for that one instrument rather than shallow for many.
Does it need write access to my account?
No. Use a read-only API key. The sync only reads fills, income and funding. It can never place, modify or withdraw.
Do I need Docker?
It is the easiest path: app and Postgres bundled, migrations on start. You can also point it at any external Postgres.
Is the AI required?
No. The whole dashboard runs without it. Set an Anthropic key and the per-trade audit and AI Coach turn on, through your own key.
Is the AI just per-trade, or is there a coach?
Both. The per-trade audit reads a single position; the AI Coach reviews a whole week or month, grades how you kept last period's rules, and sets new ones. Both run only when you ask, through your own key, and stay in your database.
Is this a signal bot?
No. It analyzes trades you already took. Finding the setup is the radar's job, that is NextScalp.