# QuantDinger

> Self-hosted trading stack from AI market research to backtest to live execution

- Category: Agents & Automation / Finance & Market Agents
- Language: Python
- Licence: Apache-2.0
- Repository: https://github.com/OpenByteInc/QuantDinger
- Website: https://www.quantdinger.com
- Canonical page: https://ai-tldr.dev/tools/quantdinger/

## In short

QuantDinger is Self-hosted trading stack from AI market research to backtest to live execution. It is aimed at independent traders, Python strategy authors and small teams. It sits in the Finance & Market Agents category of the Agents & Automation tools and is open-source, written in Python and licensed Apache-2.0.

## Overview

QuantDinger describes itself as an open-source AI trading OS: a single self-hosted stack that carries a trading idea through the whole loop — AI market research, a Python strategy, a server-side backtest, paper trading, live execution and monitoring — instead of leaving each stage to a separate tool. It is aimed at independent traders, Python strategy authors and small teams.

The design is local-first by intent. Market data, strategy code, broker credentials and the deployment itself stay under the operator's control, and the project is explicit that it is not a black-box signal service: you write the strategy, you set the risk limits, you hold the keys. Execution reaches both crypto exchanges and traditional brokers, and the platform is reachable from the web UI, a mobile H5 view, an HTTP API, an Agent Gateway and MCP — so an LLM agent can drive research and reporting through the same surface a human uses.

Version 5 reorganised the backend around explicit runtime boundaries: the HTTP API no longer owns long-running loops, and trading, scheduling, Celery jobs and migrations each run as their own process against PostgreSQL, with separate Redis instances for cache and durable jobs. An optional observability overlay adds JSON logs, request IDs, Prometheus metrics, Grafana dashboards and alert rules, and a production overlay runs the backend as a non-root user with a read-only root filesystem and dropped capabilities.

## What it does

- Multi-provider AI market research and analysis feeding into strategy work
- Python indicators and a Strategy API for writing strategies as code you own
- Server-side backtesting and experiment workflows, run as finite Celery jobs
- Paper and live execution across crypto exchanges and traditional brokers
- Access from web, mobile H5, an HTTP API, an Agent Gateway and MCP, so agents and humans share one surface
- PostgreSQL-backed state with durable workers, audit logs and an optional Prometheus, Grafana and Alertmanager overlay
- Built-in user management, billing, payments and settlement for running a multi-tenant service

## Getting started

QuantDinger ships as a Docker Compose stack. The one-command installer pulls prebuilt images from GHCR, asks for administrator credentials, generates secrets and starts everything. Docker with Compose v2 is the only prerequisite — no Node.js or local Python environment is needed.

1. Install the stack (Linux or macOS)

   The installer prompts for the initial administrator credentials, generates the required secrets, downloads the GHCR Compose stack and starts it.

```bash
curl -fsSL https://raw.githubusercontent.com/OpenByteInc/QuantDinger/main/install.sh | bash
```

2. Install the stack (Windows)

   The PowerShell installer does the same thing on Windows.

```text
irm https://raw.githubusercontent.com/OpenByteInc/QuantDinger/main/install.ps1 | iex
```

3. Open the interfaces

   Once the containers are up, the web app, the mobile H5 view and the API health endpoint are served locally.

```text
Web:        http://127.0.0.1:8888
Mobile H5:  http://127.0.0.1:8889
API health: http://127.0.0.1:5000/api/health
```

4. Start in paper mode, then decide about live trading

   QuantDinger can submit real orders once live trading is explicitly enabled. The project's own guidance is to start with paper trading, use restricted API keys, and review the risk and compliance requirements for your jurisdiction. On a manual Docker deployment the legacy default administrator credential exists only for backwards compatibility — change it before the first start or immediately after the first login.

## When to use it

- Research a market with an LLM, turn the conclusion into a Python strategy, and backtest it without leaving one stack
- Run a strategy in paper mode against live crypto or broker data before risking capital
- Keep broker credentials, strategy code and market data on infrastructure you control rather than a hosted signal service
- Expose a trading workspace to an agent over MCP or the Agent Gateway so research and reporting can be automated