TradeCanvas is a JavaScript/TypeScript library for building trading charts directly in your web app: candlesticks, 134 indicators, 76 drawing tools, order placement on the chart, replay, watchlist, and 30 interface languages. MIT licensed, no dependency on third-party libraries, can be self-hosted, no logo watermarks, and no usage fees.
Get started in 3 steps
Installation:
npm install @tradecanvas/chart
Add a div to contain the chart in your HTML:
<div id="chart" style="height: 600px"></div>
Create your first chart with real-time data from Binance:
import { ChartWidget } from '@tradecanvas/chart/widget' import { BinanceAdapter } from '@tradecanvas/chart' new ChartWidget(document.getElementById('chart')!, { symbol: 'BTCUSDT', timeframe: '5m', theme: 'dark', adapter: new BinanceAdapter(), })
Now you have a complete trading interface: toolbar, indicators, drawing tools, alerts, and replay. If you only need a pure chart to build your own UI around, use new Chart(...) and then call chart.setData(bars). There are wrappers available for React, Vue, and Svelte. If you want to use your own data (from a brokerage API, internal WebSocket), you need to write an adapter with two parts: fetching historical data and receiving ticks.
Detailed guide: https://tradecanvas.dev/vi/docs/getting-started/
What AI agents can do with TradeCanvas
When you ask a coding agent (Claude Code, Cursor, Codex, Copilot) to “add a BTC chart with RSI and price alerts,” the agent often misguesses the option names or indicator parameters, and such errors don't throw warnings during execution. TradeCanvas provides agents with two tools to eliminate guessing:
MCP server (@tradecanvas/mcp): agents can look directly into the library. For example, what parameters the RSI indicator takes, what its line is called, how many points each drawing tool needs, what options the chart has. Agents can also verify the option object before execution: typing theme: 'drak' or timezone instead of timeZone will raise an immediate error with a suggestion for the correct name. The MCP server is purely for lookup and validation; the coding remains the responsibility of the agent.
Skill: documentation guiding agents on how to use the library correctly.
Install in one line:
claude mcp add tradecanvas -- npx -y @tradecanvas/mcp
In tests on 24 programming tasks (each task executed twice with skill, twice without): with skill, 48/48 successful attempts, averaging 10 exchanges and 169k tokens per task; without skill, 47/48 successful attempts, 15 exchanges and 221k tokens.
Additionally, agents can read analytical data without the need to draw: detectCandlePatterns(bars) returns 64 candlestick patterns as data, along with direction (bullish/bearish) and type (reversal/continuation), and can run in Node.
Who it's suitable for
Securities companies, investment apps in Vietnam: already have regulations from HOSE, HNX, UPCoM, and VN30F derivatives (price steps, ceiling/floor prices, lots), with a Vietnamese interface. When connecting ExecutionAdapter to your order placement system, customers can drag stop-loss/take-profit directly on the chart.
Internal trading platforms for funds or prop desks: a maximum grid of 9 charts synchronized on symbol, timeframe, and crosshair, with orders and positions displayed on the chart, layout saved on your server. Fully self-hosted, with no third-party data interactions.
Testing strategies and trading bots: backtest using @tradecanvas/analytics (preview version) then draw each order on the chart, followed by a market replay with a simulated account to observe how the bot performs before deploying real funds. A complete example
Tổng hợp các bản 2.11.1 đến 2.13.0 (15–16/11/2026).
Replay
Thị trường Việt Nam
Mẫu nến không cần chart
Finance chart và chi tiết khác
Nâng cấp: npm install @tradecanvas/[email protected]
Chi tiết: https://tradecanvas.dev/vi/changelog/

Your trading kit to survive the crypto market

Dịch, lồng tiếng và chèn phụ đề cho video review phim tự động

Build AI workflows, no configuration. Type safe, chạy trên mọi backend

Chuyển đổi Truyện thành Audio

Phát hiện và khắc phục lỗ hổng mã nguồn

Cửa hàng ứng dụng MacOS của người Việt
Open charting for the trading app, accessed by agents through MCP.
Auto-translated ·No reviews yet
No one has supported TradeCanvas yet.
Reviews & comments