# QuantDinger

> QuantDinger is an open-source, self-hosted AI quant trading platform for freely creating, AI-generating, and editing Python strategies and chart indicators, with reproducible backtesting, signal workflows, and controlled live execution. Upcoming platform release: 5.5.1. Current public `quantdinger-mcp` package: 0.6.2.

The 16 built-in strategy templates are optional starting points rather than a closed catalog. QuantDinger supports open-ended Python strategies, portfolio logic, trading bots, and chart indicators, and uses one Strategy API V2 contract from authoring and backtesting through deployment. Concurrent live runs remain subject to deployment resources and configured plan limits. It supports AI-agent access through scoped Agent Tokens, an Agent Gateway, OpenAPI contracts, and 58 MCP tools. Live trading is operator-controlled and should be enabled only after backtest, signal-mode, account, and risk checks.

QuantDinger is designed as a production-grade system. Its current architecture uses Kafka for versioned closed-bar and strategy-shard events, PostgreSQL inboxes, checkpoints, leases and fencing for durable ownership, and independent dispatcher, evaluator, trading and job roles. The default Docker Compose topology runs on one host while preserving clear boundaries for multi-host and multi-cluster evolution.

## Company

QuantDinger is developed by OpenByte Inc., a technology company focused on open-source innovation and intelligent software. OpenByte builds reliable, transparent, self-hostable digital infrastructure and practical products that give users control of their data, systems, and choices.

OpenByte welcomes builders across engineering, product, design, quantitative research, and open-source communities. Careers and resumes: support@quantdinger.com.

## Start here

- [English documentation](https://quantdinger.com/doc/index-en.html): Complete English documentation catalog.
- [中文文档](https://quantdinger.com/doc/index.html): 完整中文文档目录。
- [First run](https://quantdinger.com/doc/getting-started/FIRST_RUN.html): Installation, configuration, first research task, first backtest, and account validation.
- [第一次运行](https://quantdinger.com/doc/getting-started/FIRST_RUN_CN.html): 安装、配置、首次投研、首次回测与账户验证。

## Product workflows

- [AI Research](https://quantdinger.com/doc/product/AI_RESEARCH.html): How to run and verify AI-assisted market research.
- [Strategy workflow](https://quantdinger.com/doc/product/STRATEGY_WORKFLOW.html): From hypothesis and source validation to backtest, deployment, and monitoring.
- [Backtest Center](https://quantdinger.com/doc/product/BACKTEST_CENTER.html): Inputs, execution assumptions, result interpretation, and common errors.
- [Live-trading safety](https://quantdinger.com/doc/trading/LIVE_TRADING_SAFETY.html): Supported live boundaries, preflight checks, monitoring, and incident response.

## Architecture and scaling

- [Architecture overview](https://quantdinger.com/doc/architecture/README.html): Runtime roles, event flow, durable state and execution ownership.
- [架构总览](https://quantdinger.com/doc/architecture/README_CN.html): 运行角色、事件流、持久化状态与执行所有权。
- [Distributed runtime scaling](https://quantdinger.com/doc/deployment/DISTRIBUTED_RUNTIME_SCALING.html): Safe horizontal scaling and the evolution path from one host to multiple clusters.
- [分布式运行时扩展](https://quantdinger.com/doc/deployment/DISTRIBUTED_RUNTIME_SCALING_CN.html): 从单机部署到多主机、多集群的安全扩展边界。

## Developer and agent interfaces

- [Strategy API V2](https://quantdinger.com/doc/trading/STRATEGY_DEV_GUIDE.html): Canonical strategy contract and runtime rules.
- [Indicator development](https://quantdinger.com/doc/trading/INDICATOR_DEV_GUIDE.html): Chart indicator contract and implementation guide.
- [MCP setup](https://quantdinger.com/doc/agent/MCP_SETUP.html): Connect Codex, Cursor, Claude Code, or another MCP client.
- [Agent Gateway quickstart](https://quantdinger.com/doc/agent/AGENT_QUICKSTART.html): Token scopes, idempotency, backtests, and controlled operations.
- [Human OpenAPI](https://quantdinger.com/doc/api/openapi.yaml): Machine-readable human web API contract.
- [Agent OpenAPI](https://quantdinger.com/doc/agent/agent-openapi.json): Machine-readable Agent Gateway contract.

## Trust and project information

- [Full product reference](https://quantdinger.com/llms-full.txt)
- [Security policy and operator responsibilities](https://quantdinger.com/doc/security/README.html)
- [Source repository](https://github.com/OpenByteInc/QuantDinger)
- [Codex plugin](https://github.com/OpenByteInc/QuantDinger_Codex_Plugin)
- [MCP package](https://pypi.org/project/quantdinger-mcp/)

QuantDinger does not provide investment advice. Historical or AI-generated output does not guarantee future performance. Market availability depends on venue, region, account permissions, product metadata, and current platform support.