Does starting the dashboard risk real money?

No. The default portfolio is simulated in SQLite with no exchange keys. Live or testnet execution is a separate documented path, and market streaming alone never moves positions.

How does the autonomous agent turn ticks into trades?

Every tick is evaluated across active timeframes with structured judgments, while risk settings, volatility, and confidence thresholds gate allocation and automatic take-profit and stop-loss exits.

Can I deploy the whole stack on Vercel alone?

The dashboard can host on Vercel, but the Python collector owns the websocket, stream, portfolio, and log files. Run it on an always-on host and point the dashboard to its URL with CORS enabled.

Does Jev execute orders or choose position sizes itself?

It is typed judgments only. The model does not place orders directly or pick arbitrary sizes; Python enforces cash, position-percentage, allocation, and confidence limits before any paper trade.

What do I need installed, and which assets can I actually trade?

Node.js 20 or newer, Python 3.12 recommended, and a TypeSafe API key. The feed subscribes to seven assets: ADA-USD, XRP-USD, ETH-USD, BTC-USD, SOL-USD, BNB-USD and TRX-USD. Install into the pinned root .venv rather than the older pipeline venv, run npm install, and put the key in pipeline/.env under either TYPESAFE_AI_API_KEY or TYPESAFE_API_KEY.

Is it safe to run one instance that several people share?

The README warns against it for a shared public trading account: the Python service has one shared in-memory paper portfolio and one active key for all connected dashboard users. A visitor can enter their own key in the optional TYPESAFE KEY field; it is sent to the feed over its config endpoint and kept in that process memory only, not stored in the browser or written to agent logs.

What decides position size, and what triggers a take-profit or stop-loss exit?

Python enforces the hard limits: shared available cash, the maximum wallet position percentage you set, the risk-profile allocation multiplier, and the confidence threshold for buy and sell actions. You pick Conservative, Balanced or Aggressive, which sets that threshold and multiplier. Take-profit and stop-loss levels come from Jev's analysis plus volatility measured by ATR and your stated risk appetite.

CommunityJev Tools & CLIs

Jev-Trades

Next.js dashboard plus Python feed that streams market data, computes indicators across timeframes, sends enabled states to Jev, and applies the judgments to a SQLite paper portfolio.

See highlights

Open github.com

Stars: 48
Language: Python
License: APACHE-2.0
Published by: Community
Author: zadescoxp
Last verified: Oct 2026

zadescoxp/Jev-Trades

User repository on GitHub

View on GitHub

Trading bot with the all new TypeSafe AI's first system one model named as Jev

Forks: 16
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.Homepage

About this project

A Next.js dashboard for live crypto market data and TypeSafe-powered trading. The system streams live market data from Yahoo Finance through yfinance (default), calculates technical indicators across multiple timeframes, sends explicitly enabled trading states to TypeSafe, and applies the returned decisions to a portfolio powered by a local SQLite database. It features both autonomous trading by the Jev agent and a fully-featured manual trade desk with automatic Take Profit and Stop Loss execution.

Quoted from the project's README. Read the full README.

Tagscommunity jevbest catalog-import awesomejev

Works forJev for Classification

Used inJev for Software & DevTools

Languages: Python

Highlights

Live Yahoo Finance feed with selectable 1-minute to 4-hour timeframes and per-timeframe indicator panels.
Jev analyzes each tick across active timeframes and returns structured judgments for the paper ledger.
Manual desk supports buys, sells, custom take-profit and stop-loss, plus agent-decision and trade logs.
Two-process run: Python collector on local stream and config ports, Next.js dashboard with asset and risk controls.

Watch out

The default is simulated paper trading, and the Python feed plus local storage are not suited to multi-instance production without a database or pinned worker. Copy reviewed against the repository README on 2026-10-04, not from running it.

Questions about this entry

Short answers drawn from the linked project or page.

More Jev resources like this

jevclip

Video tool where Jev scores subtitle segments and code assembles a short highlights cut plus a de-watered full cut with timestamped summaries.