Vibe-Trading
An open-source Python agent framework for quantitative trading, backtesting, and portfolio aggregation via MCP.
Connect market data feeds, run backtests, track shadow accounts, and integrate brokers via MCP.
Vibe-Trading is an open-source Python framework designed for automated trading strategies and portfolio management using AI agents. It connects market data loaders, backtesting engines, and broker APIs through Model Context Protocol (MCP) standards. It suits quantitative developers and financial engineers building agentic trading systems.
What it is
The tool executes multi-agent workflows that ingest market data, validate backtest metrics through strict evidence gates, and manage broker connections. It provides portfolio aggregation, shadow account tracking, diagnostic run cards, and fail-closed parsers for broker statements and CSV exports.
Who it's for
It suits Python developers, quantitative analysts, and financial engineers who require programmatic control over autonomous trading pipelines. It does not suit non-technical investors looking for a simple retail trading application or graphical interface.
Key features
Model Context Protocol Integration
Connects broker tools and financial data feeds using standardized MCP connectors.
Strict Evidence Gating
Validates strategy backtest results before allowing agents to adapt or execute trading signals.
Multi-Source Portfolio Aggregation
Aggregates balances across multiple broker accounts while refusing malformed or unmapped data.
Paper and Shadow Account Simulation
Simulates orders locally against real account snapshots without risking live trading capital.
Up and running in 3 steps
Our honest verdict
What's great
- Built-in evidence gates block unverified or flawed backtest results from live agent adaptation.
- Supports standardized Model Context Protocol (MCP) connectors for modular broker and data integrations.
- Includes strict, fail-closed parsers for portfolio exports and market data pagination.
- Offers broad multi-broker support including Alpaca, Zerodha, Dhan, Shoonya, and IBKR.
Where it falls short
- Requires significant Python development experience and self-hosted infrastructure to maintain.
- Demands active management of broker API keys, data provider credentials, and execution environments.
- Strict fail-closed data rules can reject custom data connectors if schema formats do not match exactly.
Pricing & license
Vibe-Trading is free under the MIT licence. You are responsible for all self-hosting server costs, data provider API subscriptions, and broker execution fees.
Open remote AI-training roles
880 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Does Vibe-Trading issue its own cryptocurrency token?
No, the repository maintainers explicitly confirm they have never launched or endorsed any cryptocurrency token or memecoin.
Can I use Vibe-Trading with live broker accounts?
Yes, it supports live broker connectors like IBKR and Alpaca alongside local paper trading connectors.
