Trading Journal Software

Two numbers decide this purchase and neither is the monthly price. The first is whether your broker is on the short list a journal can connect to and sync by itself, rather than the list of file formats it can parse — every vendor here quotes the second number in the hundreds. The second is which asset classes are first-class, because futures, options and forex each need arithmetic a stocks-only journal does not have. The mainstream tier runs roughly $15 to $50 a month.

• Free with no trade cap, if you type the trades in — StonkJournal, which groups multi-leg options and charges $10 a month only for file import and its AI coach; or TradeBench, free since 2011 in exchange for one sponsor click a month, which sizes each trade from your risk rules before entry and does not take options.
• A grade for keeping your own rules, not for the P&L — TradeJARnal, from $9.99 a month, typed in by hand, grades each trade on the rules you ticked as kept; 14 days of Pro without a card, and no refunds after that.
• The largest free tier that imports — TradesViz, 3,000 executions a month, stocks only. Its paid tiers, from $19.99, add options with spread detection and a payoff chart.
• Options spreads grouped as one position — check the tier, not the pitch. TraderSync leads with options, but its Auto Spread Detection is an Elite-only beta at $79.95 a month, which the support pages put at about 80% accurate. TickerReplay does it on every plan, free included, and links rolls and assignments too, but takes stocks and options only; paid from $19 a month.
• One price, everything included, no tiers — Edgewonk at $197 a year, a subscription that ends your access when you cancel; or the two you buy outright: TradingDiary Pro at $149 once, Windows only, with $49 a year optional after the first year for newer versions, and Greenwick Journal at $49.99 once, a stocks-only journal that runs offline in a browser on any OS, with no trial.
• US stocks and options, cheapest paid tier — Chartlog from $14.99.
• A journal that lives on the phone — Mantis, iPhone only, with an emotion tag on every trade and P/L by tag on Premium; free for 10 trades a month, $4.99 a month after, and nothing in the cloud, which also means no second device and no backup beyond a CSV export.
• A journal and a portfolio tracker on one account — Trademetria.
• Open source, on your own machine — TradeNote, GPL-3.0 and the only self-hosted journal here; broker files are uploaded, never synced.

Two jobs, and a boundary worth getting right

Journaling is capture: every execution, the screenshot, the note written while it still stings, the tag that says which setup it was and which rule you broke. It is judged on import fidelity and on friction, because a journal you stop filling in at the end of a bad week is worth nothing at all.

Analytics is what turns that record into an answer. It is judged on whether the answer is something you can do differently tomorrow — "you lose money in the first fifteen minutes", "your losers are twice the size of your planned risk" — rather than another dashboard.

The boundary with a portfolio tracker is the one that costs money to get wrong, because both products show a rising line and both call it performance. A tracker is built around positions: what you hold, what it cost, what it is worth, and what the revenue authority will want to know. A journal is built around executions: the order of them, the minute of them, and the three scale-ins a tracker quietly nets into one line. Ask which question you are actually paying to answer. Vendors happily sell both — Trademetria ships a journal and a portfolio tracker as separate products on one account — but almost nothing does both jobs well, and a journal is the wrong place to look for a tax report.

What actually decides the choice

"Supported brokers" is two lists wearing one number. Every vendor in this category quotes a count in the hundreds; that count is file formats it can parse. The list that matters is the short one — brokers it can connect to and pull from on its own. TradeZella's marketing says 500-plus brokers and platforms; take out the hundreds of individual MetaTrader brokerages that all reach it through the same MetaTrader connection and roughly forty named brokers and platforms remain, about fifteen of which sync automatically. TraderSync's table is the most explicit: 76 of its 491 names carry an Autosync mark, and the mark's own tooltip says to assume everything else is a CSV. Chartlog prints "Automatic" or "Manual Upload" beside every broker on its integrations page; four of 21 say automatic, three of those four are PropReports feeds for direct-access prop brokers most people have never had an account with, and the fourth is TD Ameritrade, whose accounts Schwab finished moving in May 2024. The same broker often appears on both lists at a different level of support. Check yours by name, and check which word is next to it.

There are three import mechanisms and they fail differently. A live connection through an aggregator authorises read-only access and re-syncs on a timer — TradesViz's native connections pull every 24 hours — and it breaks when the broker rotates its API or an authorisation expires. A statement or CSV parser reads a file you export yourself, and it breaks silently the day the broker redesigns the export: columns move, a new fee line appears, and the importer either errors or, worse, imports the wrong number. Manual entry never breaks and never scales. Most non-US brokers, most of MetaTrader and nearly all prop-firm platforms land in the second or third bucket.

Kinfo is built on the first mechanism alone, on purpose. About twenty US brokers sync read-only overnight, so trades appear the next morning, and there is no manual entry, no file upload and no way to exclude a trade from the numbers. That is the product: the record is the broker's rather than yours, which is what lets it back a public verified profile. It also means an unsupported broker gets no journal at all. The free plan syncs and computes basic statistics; notes, tags and the full history are Pro, at $24.90 a month.

Instruments are a hard boundary, not a checkbox. Chartlog covers US stocks and options and stops there. TradeBench takes stocks, futures, forex and crypto and no options at all. TradesViz's free tier is stocks only, with the other six asset types on Pro. Tradervue takes futures only from a published whitelist, contract by contract, and anything not on it has to be requested. The reason is structural: futures need a contract multiplier and a tick value per symbol before P&L is even arithmetic, options need a multi-leg spread treated as one position with one result, and forex needs pip values and a second currency. A journal that "supports" futures by letting you type the profit into a box is a spreadsheet with better fonts. If you trade more than one asset class, this eliminates more of the listing below than price does.

Executions are not trades, and the journal decides which is which. Somewhere in the import a rule turns a stream of fills into round trips — usually FIFO, usually splitting when the position goes flat, sometimes merging everything in one symbol on one day, with a separate rule for options spreads. Your trade count, win rate and average holding time are outputs of that rule, not facts about your trading, and two journals fed the identical file will disagree about all three. Find out whether the rule is visible and adjustable before you trust a number derived from it. TraderSync's two ordinary grouping modes, Normal and Split When Possible, work by symbol and time window and know nothing about spreads. The rule even decides how much of a free plan you have used: Tradervue's 30-trade monthly quota counts merged trades, not executions.

Gross P&L is usually the default, and it is not your P&L. Tradervue sells commission and fee support as a Gold feature, at $49.95 a month, so net P&L is not on Silver. TradeZella's sync fills in commissions with each trade; a file upload carries them only if the broker's export does. For anyone trading size or frequency, fees are the entire distance between an edge and an expensive hobby, so a journal that cannot see them is analysing a fiction.

Prop-firm accounts change what a "broker connection" even is. A funded trader has no broker account to connect: the route in is whatever platform the firm runs — Tradovate, Rithmic, NinjaTrader, cTrader, DXtrade, Match-Trader, or a firm-branded platform like Topstep's TopstepX — and the firm can change that platform out from under you, which retires your integration with it. The numbers that decide a funded account are the firm's rules rather than P&L: daily loss limit, trailing drawdown, profit target, consistency. Several journals now track those explicitly; a journal that reports only profit is missing the one number that ends the account.

The analytics you will actually use are the ones you have to feed. Six hundred statistics is a boast. The reports that change behaviour are the ones built on something you typed — the setup name, the mistake tag, whether you followed the plan, where your stop was — because an import can tell a winner from a loser and cannot tell a good trade from a lucky one. That is the real price of this category: a few minutes at the end of every trading day, indefinitely. Choose the product that makes those minutes cheapest for you, not the one with the longest stat list. The same goes for projections: TradeFuse's Monte Carlo simulator projects results forward from your current statistics, so it extends whatever the tags and the round-trip rule put into them.

What the money looks like. The mainstream tier runs roughly $15 to $50 a month: Chartlog from $14.99, TradesViz from $19.99, Tradervue and TraderSync at $29.95 with higher tiers at $49.95 and above, TradeZella from $35, or about $26 a month billed annually. The cheapest tier is not always the one that imports for you: TradeFuse starts at $29, but automated import begins on its $49 plan, and Trademetria keeps broker auto-sync for Pro alone. Edgewonk sits outside the pattern with one plan at $197 a year, everything included, and a 14-day money-back guarantee instead of a trial; VAT takes it to $234 in Germany, and on 14 September 2026 a promotion was stretching the same $197 over 16 months. Two here are bought rather than rented. TradingDiary Pro is $149 once for a Windows desktop program whose database is a single file on your own PC, which does not sync between machines, with $49 a year after the first year for new versions. Stop paying and your copy keeps working, but broker import fixes ship in new versions, and the licence still checks in with the vendor's server and turns read-only after five days without it. Greenwick Journal is $49.99 once for a stocks-only journal that runs offline in a browser on any OS, with no account, no fee after and nothing after the purchase to check in with; the journal lives in one browser on one computer and moves only by backup file. Trials are short where they exist — seven days at Chartlog, Tradervue, TraderSync, on TradesViz's paid tiers and on Mantis's $4.99 Premium, thirty at TradingDiary Pro — and TradeZella and Greenwick Journal have neither a trial nor a free plan, Greenwick offering a 14-day refund instead. Since switching means re-importing your history and re-tagging it, the annual discount is worth less than it looks until you are sure.

The vocabulary, pinned down

MFE and MAE are maximum favourable and maximum adverse excursion — how far the trade went your way before you got out, and how far it went against you before it came good. Sold under half a dozen names. They are the statistics that argue with your exits.

R and R-multiple express a result as a multiple of the risk you took: a trade that made three times your initial stop distance is +3R. It is only meaningful if the journal knows where your stop was, which means recording it before the trade, not after. An R report built on a stop you never wrote down is built on a number you invented afterwards.

Expectancy is the average result per trade, in R or in currency. It is what win rate is usually standing in for, and it is the better number: a 35% win rate can be an excellent business and an 80% win rate can be a slow bankruptcy.

"Backtesting" in a journal means replay. Bar-by-bar playback of historical data where you click entries and exits by hand, producing a sample you generated yourself. It is genuinely useful for rehearsing a discretionary setup out of market hours, and it is not what backtesting means in a systematic framework — there is no rule set, no parameter sweep and no out-of-sample. Vendors charge for it by speed and precision, so read what the tier actually buys.

"AI coach" means a language model reading your own statistics. It is only as good as the tags and notes underneath it, which loops back to the point above: the input is the work, and nothing in this category removes it. It also means your journal goes to a model provider. TradeFuseAI, a chat that answers questions about your journal with metrics and charts, starts on TradeFuse's $49 plan and sends journal data to third-party models; StonkJournal's coach sends the relevant part of the journal to OpenAI or Anthropic per question. TradingDiary Pro turns it round with a read-only MCP server on your own machine, which Claude Desktop or another MCP client can query while the app is open — though whatever it returns still goes to the AI provider as part of the conversation.