QuantDinger

Your Private AI Quant Operating System

One deployable stack for charting, AI market research, Python indicators & strategies, backtests, and live execution—on your own servers and your own keys.

Self-hosted quantitative platform: from idea and AI-assisted coding to paper-style workflows and exchange-connected live trading, with optional multi-user and billing primitives for operators.

English · 简体中文 · 日本語 · 한국어 · ไทย · Tiếng Việt · العربية

SaaS  ·  Video Demo  ·  Website  ·  AWS Marketplace

     

Quick start · Repositories · AI agents & MCP · Overview · Features · Visual tour · Architecture · Install · Docs · FAQ · License

QuantDinger is a self-hosted, local-first quantitative platform: AI-assisted research, Python-native strategies, backtesting, and live trading (crypto, IBKR stocks, MT5 forex) in one product—not a loose collection of scripts and SaaS tabs.

_{End-to-end architecture: market data feeds the five-layer engine and exits to live execution, closing the quant loop from idea to monitoring.}

Prerequisites: Docker with Compose (Docker Desktop on Windows/macOS, or Docker Engine + Compose plugin on Linux), and Git. Node.js is not required (prebuilt UI is in frontend/dist).

One line (or run the same steps separately):

git clone https://github.com/brokermr810/QuantDinger.git && cd QuantDinger && cp backend_api_python/env.example backend_api_python/.env && chmod +x scripts/generate-secret-key.sh && ./scripts/generate-secret-key.sh && docker-compose up -d --build

If ./scripts/generate-secret-key.sh fails with “Permission denied”, run chmod +x scripts/generate-secret-key.sh and retry. If docker-compose is not found, try docker compose (Compose V2).

Use PowerShell (not CMD) in a folder where you want the project. Docker Desktop must be running (WSL2 backend recommended).

git clone https://github.com/brokermr810/QuantDinger.git
Set-Location QuantDinger
Copy-Item backend_api_python\env.example -Destination backend_api_python\.env
$key = & python -c "import secrets; print(secrets.token_hex(32))" 2>$null
if (-not $key) { $key = & py -c "import secrets; print(secrets.token_hex(32))" 2>$null }
if (-not $key) { $key = & python3 -c "import secrets; print(secrets.token_hex(32))" 2>$null }
if (-not $key) { Write-Error "Install Python 3 from python.org (tick 'Add to PATH') or use Git Bash with the macOS/Linux block above." }
(Get-Content backend_api_python\.env) -replace '^SECRET_KEY=.*$', "SECRET_KEY=$key" | Set-Content backend_api_python\.env -Encoding utf8
docker-compose up -d --build

If docker-compose is not recognized, use docker compose (space, no hyphen). If Git is missing, install Git for Windows.

If you installed Git for Windows, open Git Bash and you can use the macOS / Linux one-liner above (Bash + chmod + ./scripts/generate-secret-key.sh).

Then open http://localhost:8888, sign in with quantdinger / 123456, and change the default admin password before any real use. For prerequisites, configuration details, first-run checks, and troubleshooting, continue to Installation & first-time setup below.

This monorepo ships the backend, Docker Compose stack, documentation, and a prebuilt web UI under frontend/dist. Use the sibling repos when you need source-level UI changes or the mobile app:

Note: Node.js is only required if you build the web UI from QuantDinger-Vue; the default Docker quick start does not need it.

QuantDinger ships an Agent Gateway at /api/agent/v1 and a small MCP server that wraps it as Model Context Protocol tools. Once you sign in once and issue a token, your AI client can read markets, manage strategies, run backtests, and (paper-only by default) place trades — without ever seeing your exchange keys or your admin JWT.

Two safety properties are non-negotiable: every agent call is audit-logged, and trading-class tokens are paper-only by default. Live execution requires both paper_only=false on the token AND AGENT_LIVE_TRADING_ENABLED=true on the server.

The MCP client and the wiring in Step 2 are identical for both paths — only the value of QUANTDINGER_BASE_URL changes.

Path A · Hosted (ai.quantdinger.com) — try it in 30 seconds. Sign up → open Sidebar → Agent Tokens → Issue Token. The hosted instance is locked to paper_only=true and the T (Trading) scope is rejected at issuance — agents can read markets, manage strategies in your tenant, and run backtests, but never route real-money orders. Set QUANTDINGER_BASE_URL=https://ai.quantdinger.com. Best for: trying QuantDinger from Cursor / Claude Code without installing anything; demos; research notebooks against shared datasets.

Path B · Self-hosted (this repo) — production / private data / live trading. After the Try in 2 minutes Docker bring-up, log in as admin and open Sidebar → Agent Tokens (or http://localhost:8888/#/agent-tokens). You decide scopes (incl. T), market/instrument allowlists, rate limits, and whether AGENT_LIVE_TRADING_ENABLED=true is ever flipped. Set QUANTDINGER_BASE_URL=http://localhost:8888 (or your LAN URL). Best for: anyone with their own exchange keys, anyone with private strategies/data, teams behind a VPN, or anyone who eventually wants live execution.

In either path:

1. Click Issue Token → name it (cursor-mcp, claude-research, …).
2. Pick scopes — start with R + B (read + backtest); add W to let the agent create/edit strategies.
3. Copy the token once — the dialog shows the full string once; the server only keeps a SHA-256 hash.

Prefer the CLI? See docs/agent/AGENT_QUICKSTART.md for the equivalent curl.

The MCP server lives in mcp_server/. Two transports work everywhere:

A. Local stdio (Cursor, Claude Code, Codex desktop, etc.) — the server is published on PyPI as quantdinger-mcp. Drop this into .cursor/mcp.json, ~/.config/claude/claude_desktop_config.json, or your client's equivalent (template: docs/agent/cursor-mcp.example.json):

{
  "mcpServers": {
    "quantdinger": {
      "command": "uvx",
      "args": ["quantdinger-mcp"],
      "env": {
        "QUANTDINGER_BASE_URL":    "http://localhost:8888",
        "QUANTDINGER_AGENT_TOKEN": "qd_agent_xxxxxxxx"
      }
    }
  }
}

uvx (install uv) downloads + caches the package on first run; no virtualenv setup. If you prefer pip:

pip install quantdinger-mcp
# then use {"command": "quantdinger-mcp", "args": []}

For Claude Code's CLI helper:

claude mcp add quantdinger \
  --env QUANTDINGER_BASE_URL=http://localhost:8888 \
  --env QUANTDINGER_AGENT_TOKEN=qd_agent_xxxxxxxx \
  -- uvx quantdinger-mcp

B. Remote HTTP (cloud agents like OpenClaw / NanoBot, browser IDEs, anything that can't spawn subprocesses) — run the MCP server as a long-lived service, then point clients at the URL:

QUANTDINGER_BASE_URL=https://your-host \
QUANTDINGER_AGENT_TOKEN=qd_agent_xxxxxxxx \
QUANTDINGER_MCP_TRANSPORT=streamable-http \
QUANTDINGER_MCP_HOST=0.0.0.0 \
QUANTDINGER_MCP_PORT=7800 \
quantdinger-mcp
# clients connect to http://your-host:7800

Use QUANTDINGER_MCP_TRANSPORT=sse instead for clients that only speak the older SSE transport. Put a reverse proxy in front for TLS and IP allowlisting.

Restart the IDE, then ask things like:

• "Pull the last 90 daily candles for BTC/USDT and tell me what the regime detector says."
• "Backtest the 20/60 SMA crossover on ETH/USDT 4h between 2024-01-01 and 2024-06-30 and stream the result as it runs."
• "Create a strategy named eth-trend-bot, use the indicator I just designed, leave it in stopped state."

Long-running jobs (/api/agent/v1/jobs/{id}/stream) are exposed as SSE so the agent can react to partial results without polling. Every call shows up under Agent Tokens → Audit log with route, scope class, status code, and duration.

If you're editing this repo with Cursor / Claude Code / Codex, the repo also ships a Cursor Skill at .cursor/skills/quantdinger-agent-workflow/SKILL.md that explains the Agent Gateway internals, red lines (no real keys, paper-only by default), and where to verify changes. Read docs/agent/AGENT_ENVIRONMENT_DESIGN.md for the full layered-contracts model.

Deeper links: AI Integration design · Quickstart with curl · OpenAPI 3.0 spec · MCP server README

QuantDinger is a self-hosted quantitative OS: AI-assisted research, Python-native strategies (IndicatorStrategy + ScriptStrategy), backtesting, and live trading (crypto, IBKR, MT5)—with optional multi-user roles, notifications, credits, and USDT billing. It replaces a patchwork of charts, notebooks, bots, and disconnected LLM chats with one Compose stack and your credentials in Postgres + .env.

Audience: traders and quants, Python strategy authors, small teams building internal or commercial trading products.

▶ Watch Product Demo on YouTube Click the preview card above to open the full video walkthrough.
Indicator IDE, charting, backtest, and quick trade	AI asset analysis and opportunity radar
Trading bot workspace and automation templates	Strategy live operations, performance, and monitoring

• Research & AI — Multi-LLM analysis, watchlists, analysis history; optional ensemble/calibration; NL→indicator/strategy; post-backtest AI hints; Polymarket as a research workflow. Agent Gateway + MCP for Cursor / Claude Code / Codex.
• Build — IndicatorStrategy (dataframe signals, chart overlays) and ScriptStrategy (on_bar, explicit orders); professional chart UI.
• Validate — Server-side backtests, metrics, equity curves, strategy snapshots.
• Operate — Crypto execution, quick trade, IBKR / MT5, notifications (Telegram, email, SMS, Discord, webhooks).
• Platform — Docker Compose, Postgres, Redis, OAuth, multi-user patterns, credits / membership / USDT billing toggles.

Stack: Nginx serves the prebuilt Vue app (frontend/dist); Flask API runs strategy/AI/billing services; PostgreSQL holds state; Redis backs workers. Exchanges, brokers, LLMs, and payments plug in through env-driven adapters. Crypto market data and order execution paths are separated by design.

Runtime (short): data feeds → backtest/strategy engine → live runtime → exchange adapters; pending orders dispatched per venue.

flowchart LR
    U[Trader / Operator / Researcher]

    subgraph FE[Frontend Layer]
        WEB[Vue Web App]
        NG[Nginx Delivery]
    end

    subgraph BE[Application Layer]
        API[Flask API Gateway]
        AI[AI Analysis Services]
        STRAT[Strategy and Backtest Engine]
        EXEC[Execution and Quick Trade]
        BILL[Billing and Membership]
    end

    subgraph DATA[State Layer]
        PG[(PostgreSQL 16)]
        REDIS[(Redis 7)]
        FILES[Logs and Runtime Data]
    end

    subgraph EXT[External Integrations]
        LLM[LLM Providers]
        EXCH[Crypto Exchanges]
        BROKER[IBKR / MT5]
        MARKET[Market Data / News]
        PAY[TronGrid / USDT Payment]
        NOTIFY[Telegram / Email / SMS / Webhook]
    end

    U --> WEB
    WEB --> NG --> API
    API --> AI
    API --> STRAT
    API --> EXEC
    API --> BILL

    AI --> PG
    STRAT --> PG
    EXEC --> PG
    BILL --> PG
    API --> REDIS
    API --> FILES

    AI --> LLM
    AI --> MARKET
    EXEC --> EXCH
    EXEC --> BROKER
    BILL --> PAY
    API --> NOTIFY

Fast path: Try in 2 minutes first. The steps below are the full checklist (same outcome, more detail).

This section mirrors a typical “local deploy” path: prepare the host → obtain the code → configure secrets → start the stack → verify → harden → optionally wire AI. Node.js is not required: the repo ships a prebuilt UI under frontend/dist and Nginx serves it inside the frontend container.

git clone https://github.com/brokermr810/QuantDinger.git
cd QuantDinger

cp backend_api_python/env.example backend_api_python/.env

Almost all runtime behavior is driven by backend_api_python/.env (database URL, admin user, LLM keys, workers, billing toggles, etc.). The optional repository root .env only adjusts Compose-level concerns such as ports and image mirrors (IMAGE_PREFIX).

The API refuses to start if SECRET_KEY is still the placeholder from env.example. This blocks accidental insecure deployments.

Linux / macOS (recommended):

./scripts/generate-secret-key.sh

The script overwrites the SECRET_KEY= line in backend_api_python/.env using Python’s secrets module.

Manual (any OS): generate a long random string (for example 64 hex chars) and set SECRET_KEY=... in backend_api_python/.env.

docker-compose up -d --build

Services: postgres, redis, backend, frontend (see docker-compose.yml for healthchecks and port mappings).

Default admin (change immediately in production):

• User: quantdinger
• Password: 123456 (from env.example; override with ADMIN_USER / ADMIN_PASSWORD in .env before first use if you prefer).

Also set FRONTEND_URL in backend_api_python/.env to the URL users actually use (including https:// behind a reverse proxy); it affects redirects, CORS-related settings, and some generated links.

AI analysis, NL→code, and related flows need at least one LLM provider configured. Open backend_api_python/env.example, find the AI / LLM block, copy the relevant keys into your .env (for example LLM_PROVIDER + OPENROUTER_API_KEY, or another supported provider). Restart the backend after edits.

Use Docker Desktop (WSL2 backend recommended). From PowerShell in the repo root:

git clone https://github.com/brokermr810/QuantDinger.git
cd QuantDinger
Copy-Item backend_api_python\env.example -Destination backend_api_python\.env
$key = py -c "import secrets; print(secrets.token_hex(32))"
(Get-Content backend_api_python\.env) -replace '^SECRET_KEY=.*$', "SECRET_KEY=$key" | Set-Content backend_api_python\.env -Encoding UTF8
docker-compose up -d --build

If py is not on PATH, use python or python3 in the one-liner that generates $key. Line endings should remain UTF-8; avoid editors that strip newlines from .env.

docker-compose ps
docker-compose logs -f backend
docker-compose restart backend
docker-compose up -d --build
docker-compose down

For custom ports or mirror/prefix for base images (slow Docker Hub pulls), create a file named .env in the repository root (same directory as docker-compose.yml):

FRONTEND_PORT=3000
BACKEND_PORT=127.0.0.1:5001
IMAGE_PREFIX=docker.m.daocloud.io/library/

Production-style TLS, domain, and reverse-proxy placement are covered in Cloud deployment.

After the stack is healthy: (1) run an AI asset / market analysis so LLM and data paths are verified; (2) open the Indicator IDE, load a symbol, and run a signal backtest on a small date range; (3) optionally use AI code generation to draft an indicator, then edit the Python; (4) when ready, attach exchange API keys (profile / credentials), use test connection, then explore live strategy or quick trade with execution mode you intend. This order surfaces configuration issues early before real capital.

This is the kind of Python-native strategy logic QuantDinger is designed for:

# @param sma_short int 14 Short moving average
# @param sma_long int 28 Long moving average

sma_short_period = params.get('sma_short', 14)
sma_long_period = params.get('sma_long', 28)

my_indicator_name = "Dual Moving Average Strategy"
my_indicator_description = f"SMA {sma_short_period}/{sma_long_period} crossover"

df = df.copy()
sma_short = df["close"].rolling(sma_short_period).mean()
sma_long = df["close"].rolling(sma_long_period).mean()

buy = (sma_short > sma_long) & (sma_short.shift(1) <= sma_long.shift(1))
sell = (sma_short < sma_long) & (sma_short.shift(1) >= sma_long.shift(1))

df["buy"] = buy.fillna(False).astype(bool)
df["sell"] = sell.fillna(False).astype(bool)

See full examples:

• docs/examples/dual_ma_with_params.py
• docs/examples/multi_indicator_composite.py
• docs/examples/cross_sectional_momentum_rsi.py

Polymarket is currently supported as a research and analysis workflow, not as direct in-platform live execution. It is useful for market lookup, divergence analysis, opportunity scoring, and AI-assisted review.

QuantDinger supports two main strategy authoring models:

• dataframe-based Python scripts
• buy / sell signal generation
• chart rendering and signal-style backtests
• best for research, indicator logic, and visual strategy prototyping

• event-driven on_init(ctx) / on_bar(ctx, bar) scripts
• explicit runtime control with ctx.buy(), ctx.sell(), ctx.close_position()
• best for stateful strategies, execution-oriented logic, and live alignment

For the full developer workflow, see:

• Strategy Development Guide
• Cross-Sectional Strategy Guide
• Strategy Examples

The example scripts live in docs/examples/ and are kept aligned with the current strategy development guides.

QuantDinger/
├── backend_api_python/      # Open backend source code
│   ├── app/routes/          # REST endpoints
│   ├── app/services/        # AI, trading, billing, backtest, integrations
│   ├── migrations/init.sql  # Database initialization
│   ├── env.example          # Main environment template
│   └── Dockerfile
├── frontend/                # Prebuilt web UI (sources: QuantDinger-Vue; mobile app: QuantDinger-Mobile)
│   ├── dist/
│   ├── Dockerfile
│   └── nginx.conf
├── docs/                    # Product, strategy, and deployment documentation
├── docker-compose.yml
├── LICENSE
└── TRADEMARKS.md

Use backend_api_python/env.example as the primary template. Key areas include:

Strategy: EN · CN · TW · JA · KO · Cross-sectional EN / CN · Examples

Integrations & alerts: IBKR · MT5 EN / CN · OAuth EN / CN · Telegram / Email / SMS configs under docs/ (NOTIFICATION_*).

Yes. The default deployment model is your own Docker Compose stack with your own database, Redis instance, credentials, and environment configuration.

No. Crypto is a major focus, but the platform also includes IBKR workflows for US stocks, MT5 workflows for forex, and Polymarket research support.

Yes. QuantDinger supports both dataframe-style IndicatorStrategy development and event-driven ScriptStrategy development. You can also use AI to generate a starting point and then edit it yourself.

It is both. QuantDinger is built to connect AI research, charting, strategy development, backtesting, quick trade flows, and live execution operations in one system.

The backend is licensed under Apache 2.0. The web frontend source (QuantDinger-Vue) uses a separate source-available license—review both and contact the project for commercial frontend authorization if needed. The mobile app repo is open source under its own license (see that repository).

Yes—see QuantDinger-Mobile (open source). It connects to the same backend you self-host or to SaaS.

The following links are available in-app under Profile -> Open account and may qualify users for trading-fee rebates depending on venue policies.

• Backend source code is licensed under Apache License 2.0. See LICENSE.
• This repository distributes the frontend UI here as prebuilt files for integrated deployment.
• The frontend source code is available separately at QuantDinger Frontend under the QuantDinger Frontend Source-Available License v1.0.
• Under that frontend license, non-commercial use and eligible qualified non-profit use are permitted free of charge, while commercial use requires a separate commercial license from the copyright holder.
• Trademark, branding, attribution, and watermark usage are governed separately and may not be removed or altered without permission. See TRADEMARKS.md.

For commercial licensing, frontend source access, branding authorization, or deployment support:

• Website: quantdinger.com
• Telegram: t.me/worldinbroker
• Email: support@quantdinger.com

QuantDinger is intended for lawful research, education, and compliant trading only—not for fraud, market manipulation, sanctions evasion, money laundering, or other illegal activity. Operators must follow applicable laws, licensing, and exchange rules in every jurisdiction where they deploy. This project does not provide legal, tax, investment, or regulatory advice. You use the software at your own risk; to the extent permitted by law, contributors disclaim liability for trading losses, service interruption, or regulatory enforcement arising from use or misuse.

• Contributing Guide
• Contributors
• Report Bugs / Request Features
• Email: support@quantdinger.com

Crypto donations:

0x96fa4962181bea077f8c7240efe46afbe73641a7

QuantDinger stands on top of a strong open-source ecosystem. Special thanks to projects such as:

• Flask
• Pandas
• CCXT
• yfinance
• Vue.js
• Ant Design Vue
• KLineCharts
• ECharts
• Capacitor
• bip-utils

_{If QuantDinger is useful to you, a GitHub star helps the project a lot.}