Three doors into the same journal
Ask a trader why their journal died and the answer is never “the analytics were weak.” It died at data entry. Your fills live inside Zerodha Kite or Groww; your journal lives somewhere else; and every evening someone has to carry the numbers across. If that trip takes ten minutes, it happens on winning days and gets skipped on losing ones — which is exactly backwards, because the losing days are the ones worth studying.
There are three practical routes from a Zerodha or Groww account into a journal: screenshot import, broker CSV, and manual entry. They aren't competing features — each fits a different situation, and most traders end up using all three. Here's when to reach for which.
Route 1: screenshot import — the daily default
The fastest way to log a trading day is to photograph it. After the 3:30 PM IST close, open the executed-orders screen — Kite's orders tab, or Groww's order history — take a screenshot, and upload it. AI reads the image and extracts every fill: symbol, buy or sell, quantity, price, and time. It only reports what's visible on the screen — a blurred price comes back flagged for you to fill in, not guessed.
A worked example: nine fills across two NIFTY weekly option strikes between 9:20 and 11:05. That's one screenshot, maybe two if the list scrolls. The spreadsheet alternative is forty-five cells typed by hand, each one a chance to fat-finger a strike or swap a buy for a sell. This is why screenshot import is the route for daily logging: the cost of capturing a day drops to the cost of taking a picture. The mechanics of how the extraction works are covered in From Broker Screenshot to Journal in 30 Seconds.
Route 2: broker CSV — for backfilling weeks at once
Screenshots are for today. When you need last month, use the broker's own export. Zerodha's Console has a tradebook report — pick a date range, download the file. Groww offers an order history export the same way. One file can carry hundreds of fills with exact exchange timestamps and prices; there's nothing to read or interpret, which makes CSV the natural route for bulk backfill.
The usual moment for this is week one. A journal with three days of data can't tell you anything; a journal seeded with your last two months can. Backfill May and June from the tradebook, and your first weekly review starts from a real baseline — 70-odd round trips, an actual average loss in rupees — instead of whatever you happen to remember about your own trading.
Route 3: manual entry — small, but load-bearing
Some days don't justify an upload. One stock future, in and out — manual entry is six fields and thirty seconds. It also covers the gaps the other routes can't: a smaller broker without a clean export, a fill the AI couldn't read from a cropped screenshot, or an old trade you're reconstructing from the contract note.
Treat manual entry as the exception path, not the plan. It's the slowest route per fill, and slow routes are the first thing skipped after a bad day — the failure mode that kills journals.
Between upload and journal: review, duplicates, FIFO
Whichever door the data comes through, three things happen before it counts as a trade:
- You review before anything commits. Imported fills land in a staging queue, not your journal. You see exactly what was read — symbol, side, quantity, price, time — and can correct any field before saving. Nothing enters your record without your confirmation.
- Re-importing the same screen is safe. Duplicates are detected, so uploading the same orders screenshot twice — or a screenshot that overlaps yesterday's CSV backfill — doesn't double-count anything. Practical consequence: when in doubt, import again. Snap the orders screen at lunch and again after close; the overlap is handled.
- On commit, fills become trades. Raw fills are paired FIFO into round-trip trades — an entry, its exits, one P&L. Four partial exits on one position become one decision with one outcome, which is what makes win rate and average-loss numbers mean anything. The pairing logic is explained in FIFO Round-Trips: How Your Fills Become Trades.
The only metric that matters: seconds per day
None of this is about predicting the market. A journal measures your own past decisions — and it can only measure the days that actually get logged. The honest test of any import route is what happens at 3:35 PM on a day you're down ₹9,000: if logging takes ten minutes, that day quietly never makes it in, and your data ends up flattering you.
That's the design target behind PnL Book: screenshot for the daily habit, tradebook CSV for the backfill, manual entry for the stragglers — every route ending in the same review queue and the same FIFO-paired trades. Get a full day logged in under a minute and the losing days get logged too. Those are the ones the whole exercise is for.