DEV.co
AI Frameworks · HKUDS

Vibe-Trading

Vibe-Trading is a Python-based AI agent framework for algorithmic trading that combines LLM capabilities with multi-agent orchestration, backtesting, and quantitative finance tools. It provides a unified command-line and API interface to build, test, and deploy trading strategies with integrated broker connectivity.

Source: GitHub — github.com/HKUDS/Vibe-Trading
18.5k
GitHub stars
3.1k
Forks
Python
Primary language
MIT
License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

FieldValue
RepositoryHKUDS/Vibe-Trading
OwnerHKUDS
Primary languagePython
LicenseMIT — OSI-approved
Stars18.5k
Forks3.1k
Open issues17
Latest releasev0.1.10 (2026-06-19)
Last updated2026-07-07
Sourcehttps://github.com/HKUDS/Vibe-Trading

What Vibe-Trading is

Built on FastAPI (backend) and React 19 (frontend), Vibe-Trading offers an MCP-compatible agent framework for trading logic, supports Python 3.11+, and includes backtesting, shadow account simulation, and real-time order execution capabilities. Recent commits show active refactoring toward modular API routes, UTC timestamp standardization, and performance optimization via NumPy/bottleneck fast paths.

Quickstart

Get the Vibe-Trading source

Clone the repository and explore it locally.

terminalbash
git clone https://github.com/HKUDS/Vibe-Trading.gitcd Vibe-Trading# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Quantitative Strategy Development & Backtesting

Teams building and validating algorithmic trading strategies can use the built-in backtesting framework and shadow account simulation to iterate without real capital exposure before live deployment.

Multi-Agent Orchestration for Trading Workflows

Organizations needing coordinated decision-making across market analysis, risk assessment, and order execution can leverage the multi-agent architecture and MCP tool integration for complex workflows.

LLM-Augmented Trading Research & Execution

Fintech platforms seeking to combine natural language interfaces with trading logic can integrate Vibe-Trading's agent framework to allow users to express trading intent in prose and have the LLM orchestrate data fetching, analysis, and order placement.

Implementation considerations

  • Verify Python 3.11+ environment and ensure compatible FastAPI/React 19 deployments; recent Windows baseline work shows platform-specific test isolation needed.
  • Plan for LLM integration (LangChain/LangGraph dependency management with pinned floors); understand MCP tool registration for broker and market data providers.
  • Review the modularization-in-progress state (API routes being refactored into slices as of July 2026); expect minor breaking changes in minor versions.
  • Configure broker credentials via multi-level .env search order (~/.vibe-trading/.env → agent/.env → $CWD/.env); SSRF guards on channel media and OAuth cache isolation add complexity.
  • Evaluate shadow account setup and timezone handling (UTC standardization completed July 2026); timezone-aware fixtures required for session/API consistency.

When to avoid it — and what to weigh

  • Strict Production Compliance & Audit Requirements — Early-stage projects (v0.1.10) with active breaking changes, still-open API modularization (#331), and limited security audit trail may not meet regulatory expectations for mission-critical financial systems without substantial hardening.
  • No Tolerance for Dependency Churn — Recent commits show aggressive refactoring, floor-pinning on LangChain/LangGraph/Pillow, and platform-specific fixes (Windows baseline, proxy bypasses). Teams on strict dependency lockdown should evaluate stability first.
  • Real-Time Sub-Second Trading Latency — The architecture is designed for agent-driven analysis and placement rather than microsecond-level execution; do not expect ultra-low-latency HFT performance from an LLM-orchestrated system.
  • Single-Vendor Lock-In Avoidance — Vibe-Trading integrates tightly with broker APIs (Robinhood MCP noted in recent commits) and proprietary data sources (Alibaba Cloud IQS fallback for CN); vendor independence is not a design goal.

License & commercial use

MIT License. Permissive open-source license allowing commercial use, modification, and distribution under standard OSI terms with attribution.

MIT permits commercial use without royalty or vendor approval. However, verify that your use of integrated broker APIs, market data, and LLM services complies with those vendors' commercial terms separately. The Vibe-Trading library itself has no commercial license restrictions, but trading operations and data access may.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

SignalAssessment
MaintenanceActive
DocumentationAdequate
License clarityClear
Deployment complexityHigh
DEV.co fitGood
Assessment confidenceHigh
Security considerations

Recent SSRF defenses added (July 2026) to reject CGNAT/mesh targets and QQ media redirects before fetch; OAuth cache platform-aware. Subprocess environment for backtests allowlisted instead of inheriting parent secrets. No full security audit claimed. Broker credential handling via .env multi-level search; ensure secrets not logged or exposed via MCP. Preflight provider checks now reject redirects (hardened July 2026). No claimed penetration testing or vulnerability disclosure program noted.

Alternatives to consider

QuantConnect

Established cloud-hosted backtesting and live trading platform with broader broker support, higher regulatory maturity, and institutional adoption. Trade-off: higher cost, less customizable agent logic, less control over LLM integration.

Backtrader

Lightweight, mature Python backtesting library with large community and minimal dependencies. Trade-off: no built-in agent orchestration, LLM integration, or broker-agnostic API server; requires more manual plumbing.

MLflow / Weights & Biases (for experiment tracking) + proprietary trading stack

Decoupled experiment management and ML ops for teams building custom trading systems without LLM dependency. Trade-off: no out-of-the-box trading orchestration; more greenfield engineering required.

Software development agency

Build on Vibe-Trading with DEV.co software developers

Contact our engineering team to assess integration feasibility, security posture, and roadmap alignment with your trading and quantitative finance goals.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

Related on DEV.co

Explore the category and the services that help you build with it.

Vibe-Trading FAQ

Is Vibe-Trading production-ready for live trading?
Version 0.1.10 is pre-1.0 and shows active refactoring (API modularization in progress, recent SSRF hardening, Windows baseline fixes). Test thoroughly in shadow accounts before deploying real capital. No formal security audit noted.
What LLM providers does it support?
Uses LangChain/LangGraph (pinned dependencies as of July 2026), which typically support OpenAI, Anthropic, and other compatible providers. Specific provider list not stated in README; check docs and source for definitive support matrix.
How do I integrate my broker?
MCP tool registration and provider preflight configuration required. Robinhood MCP explicitly mentioned in recent commits; others require custom MCP wrappers or direct API binding. Review provider loader and .env search order documentation.
Can I deploy this on-premises or air-gapped?
Requires Python 3.11+, FastAPI, React 19, and LLM/market-data connectivity. Air-gapped deployments would need offline LLM or local model fallback and cached market data; not designed for disconnected scenarios.

Work with a software development agency

From first prototype to production, DEV.co delivers software development services around tools like Vibe-Trading. Our software development agency staffs experienced software developers and web developers for custom software development, web development, integrations, and ongoing support across ai frameworks and beyond.

Ready to Evaluate Vibe-Trading for Your Fintech Stack?

Contact our engineering team to assess integration feasibility, security posture, and roadmap alignment with your trading and quantitative finance goals.