trading-terminal

Literally draw where you think the market is going.Get a fully managed trade back.

An AI trading terminal for Hyperliquid. Sketch your read straight onto the chart; the agent turns it into a real setup with entry, stop and target, ranks more setups beside it, and writes the Freqtrade strategy that trades the one you pick — refusing the unsafe ones on the way.

Run it yourself, or use the hosted build at terminal.superior.trade.

You draw the path. It comes back with a plan, a rating, and a deploy button.

[!TIP] Have an agent instead of a mouse? The terminal is for humans. If you want your agent trading directly — no UI, no clone — hand it one line: Read https://superior.trade/SKILL.md and register. Works from Claude Code, OpenClaw, Codex and Cursor. Pointing a coding agent at this repository instead? AGENTS.md.

News

• 2026-08-21 — A fresh clone now runs without TradingView: the bundled Lightweight Charts preview boots first, Advanced Charts drops in later (npm run setup:charts).
• 2026-08 — The validator's rule set and repair loop are written up in docs/strategy-pipeline.md. Every rule exists because a strategy once reached production without it.

The loop

1. You draw a level and ask.

"HYPE keeps rejecting off this line. Is there a mean-reversion trade here on the 15m?"

2. The agent reads your actual chart — your drawings, the visible range, the indicators you have plotted, and a screenshot of it — and answers with a plan:

3. You say deploy. It writes the strategy, and the validator reads it before anything else does:

def populate_entry_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
    dataframe.loc[
        (
            (dataframe['volume'] > 0) &                      # no signals on dead candles
            (dataframe['close'] < dataframe['bb_lower']) &
            (dataframe['rsi'] < 40) &
            (dataframe['close'] > dataframe['ema_50'])
        ),
        'enter_long'
    ] = 1

4. It runs. Live PnL, an auto-stop timer if the edge is time-boxed, and an inbox that tells you when something happened.

The other half — what you have deployed, what state it is in, and the controls.

Hosted or self-hosted

The hosted build at terminal.superior.trade is this repository with a login in front of it. Both talk to the same Superior Trade API and trade the same accounts, so the choice is about who runs the process, not about what you get.

Running your own copy means your keys stay on your machine and every prompt and safety rule is yours to change. It also means no login (see SECURITY.md), withdrawals that stop one hop short (see docs/withdrawals.md), and getting your own TradingView access before the chart will build.

What actually stops you losing money

The interesting part of this repo is not that a model writes Python. It is what happens between "the model wrote it" and "it is trading your funds".

Every generated strategy is parsed and checked before it can be submitted. A rejection is not an error you see — the complaints go back to the model, which rewrites the strategy, and the loop repeats until it passes. Some of what gets caught:

Most of these are scars: they exist because a strategy reached production without them. docs/strategy-pipeline.md has the full set and explains the repair loop.

Features

Run it yourself

Two keys. Nothing else to sign up for.

git clone https://github.com/Superior-Trade/trading-terminal.git
cd trading-terminal
npm install
cp .env.example .env.local     # add the two keys below
npm run dev                    # → http://localhost:3200

No database to provision, no login screen. An embedded Postgres writes to ./.data/ and migrates itself on first boot.

.env.example documents 38 variables. Those two are the only ones without a working default.

[!NOTE] It starts on the preview chart. TradingView's Advanced Charts is free but not redistributable, so it is not in this repository and a fresh clone runs Lightweight Charts instead — Apache-2.0, bundled, nothing to request.

Everything works on it, including drawing your read: a small toolbar on the chart draws freehand brush strokes, trendlines, rays, horizontal levels, vertical lines, rectangles and fib retracements, and the agent reads them the same way it reads TradingView drawings. What the preview still lacks is text notes and indicator studies. To get those, request Advanced Charts at tradingview.com/advanced-charts (free, a day or two) and run npm run setup:charts. The build tells you which chart it picked. The hosted build at terminal.superior.trade already runs the full TradingView chart, drawings and all. Full comparison: docs/charting-library.md.

Money

This places real orders with real funds by design. The validator above catches the mistakes we know about, not the ones we don't — so read what the agent writes before you deploy it, and start on testnet (NEXT_PUBLIC_HL_NETWORK=testnet) or with an amount you would shrug at.

A backtest is not a prediction, and neither is the agent's reasoning. Not investment advice, no promise of profit, and you are responsible for what you run.

Documentation

Development

npm run dev           # dev server on :3200
npm test              # unit tests
npm run test:e2e      # 19 checks against a running server, nothing mocked
npm run check-types
npm run lint
npm run preview       # render README.md the way GitHub will, images and all

npm run test:e2e drives the whole app — boot, database, the Superior Trade API, market data, real model calls. It never spends money: deploying and withdrawi