chart-img-mcp
Enables AI agents to capture TradingView chart snapshots via the CHART-IMG API, supporting multiple chart types and resolutions with automatic dimension clamping and authentication.
README
chart-img-mcp
A Model Context Protocol (stdio) server for CHART-IMG — TradingView chart snapshots via https://api.chart-img.com. API reference: doc.chart-img.com.
Existing MCPs?
Public MCP listings include TradingView chart servers that drive a browser (e.g. Selenium) or QuickChart / ECharts — they are not the official CHART-IMG REST API. This package wraps chart-img.com similarly to how binance-mcp wraps the Binance SDK.
Prerequisites
- Node.js 20+
- pnpm
Install
cd chart-img-mcp
pnpm install
cp .env.example .env
# Set CHART_IMG_API_KEY in .env (from chart-img.com)
pnpm run build
Run (stdio)
pnpm start
# or
node dist/index.js
Cursor / IDE
Add a server entry (do not commit real keys):
{
"mcpServers": {
"chart-img": {
"command": "node",
"args": ["/absolute/path/to/auto_trade/chart-img-mcp/dist/index.js"],
"env": {
"CHART_IMG_API_KEY": "your-api-key"
}
}
}
}
Dev without building:
"command": "pnpm",
"args": ["exec", "tsx", "src/index.ts"],
"cwd": "/absolute/path/to/auto_trade/chart-img-mcp"
MCP resources
| URI | Purpose |
|---|---|
chart-img://docs/quick-reference |
Curated agent notes (resolution limits, auth, symbols, v3 caveats, errors) with links into doc.chart-img.com. Fetched via Resources: List / Read in MCP clients. |
Static snippets stay accurate for common bumps; the live manual remains the source of truth for full parameter lists and indicator/drawing schemas.
Tools
| Name | Notes |
|---|---|
chartimg_snapshot |
Agent-friendly: v2 advanced chart → storage URL with safe defaults (symbol + interval + Volume + max dimensions). Prefer this for typical “show me this chart” requests. |
chartimg_v1_mini_chart_storage |
GET v1 mini chart → storage JSON (url, etc.) |
chartimg_v1_advanced_chart_storage |
GET v1 advanced chart → storage JSON |
chartimg_v2_advanced_chart_storage |
POST v2 advanced chart → storage JSON (preferred for large images). Width/height are clamped automatically. |
chartimg_v2_advanced_chart_image |
POST v2 advanced chart → base64 PNG/JPEG (can be large). Width/height are clamped automatically. |
chartimg_v2_layout_chart_storage |
POST v2 shared layout → storage JSON |
chartimg_v2_layout_chart_image |
POST v2 shared layout → base64 |
chartimg_v3_exchange_list |
GET v3 exchange id list |
chartimg_v3_exchange_symbols |
GET v3 symbols for an exchange (e.g. binance). Payload may be empty on some plans — use known TradingView symbols (e.g. BINANCE:BTCUSDT.P) from docs instead of relying on this list. |
Resolution: v2 chart requests use CHART_IMG_MAX_WIDTH / CHART_IMG_MAX_HEIGHT (default 800×600) to clamp width/height before calling the API, avoiding HTTP 403 when the plan’s max resolution is exceeded. When clamping happens, the JSON response includes _chartimgMcp.dimensionClamps.
Optional v2 headers: tradingviewSessionId + tradingviewSessionIdSign for TradingView subscription data (see CHART-IMG docs).
If CHART_IMG_API_KEY is unset, only chartimg_tools_disabled is registered with a short hint.
Auth (per CHART-IMG)
- v1:
Authorization: Bearer <key>(this server uses the header). - v2 / v3:
x-api-key: <key>.
Limits
Rate limits, max image dimensions, and storage TTL depend on your CHART-IMG plan — see Request limits in the official docs.
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。