Project Overview
Trading Journal AI is a local-first, open-source trading journal with an AI coach in development. It helps you organize your trades so they're easy to find, tag, and reflect on — and it turns that daily review into a habit instead of a chore.
If you've ever kept notes in a spreadsheet, a broker platform, or a pile of screenshots and then never looked at them again, this is built for you.
What it does
The app is built around one operating loop for the trading day:
Dashboard prompts. Journal captures. Coach synthesizes. Dashboard reminds.
- Dashboard prompts (in development) — during the session, the dashboard is the active-day surface: your morning plan, the rules and risk guardrail kept visible while you trade, and midday check-ins that ask "should I keep trading?" — with your answers routed into the day's recap.
- Journal captures — the durable record, and the source of truth. Import your trades, write a short recap of the day, and tag each trade in your own language (what the setup was, how you executed, how you felt). Available today.
- Coach synthesizes (in development) — an AI reviewer that reads that record against the rules you define and drafts feedback in your voice: what worked, what broke, and one focused thing to try next session. You always edit before it saves. An early version runs today with your own OpenAI key.
- Dashboard reminds (in development) — before the next session, the dashboard brings the active lesson back: the current experiment, the risk guardrail, the cue worth keeping in view.
The loop starts and ends with the journal, where your notes are a first-class citizen — written at the day level, tagged in your own words. Your trades, charts, calendar, and analytics aren't the story on their own; they're the evidence that flows through the loop so each day's review stays grounded in what actually happened. As the system matures, analytics becomes the loop's investigation layer (test whether a coach observation holds up over a bigger sample) and the playbook becomes its standards layer — the rules and setups you've approved.
What makes it different
- Your data stays on your machine. Everything lives in a local SQLite file in your project folder. There's no hosted account, and nothing is uploaded unless you choose to — by deploying it yourself, or by asking the AI coach for a review, which sends that session to the model provider you configured under your own key.
- No signup, no subscription. It's a personal tool you run locally, not a service you rent.
- Open source (MIT). Download it, fork it, run it, or use it as the starting point for your own journal.
Who it's for
Active traders who want a private, structured review habit — especially anyone who cares about keeping sensitive account history off someone else's server. You don't need to be a developer to use it, though you will run a couple of commands to install it the first time.