QuantCoder 2.0.0

Local-first CLI for generating QuantConnect trading algorithms from research papers — powered by Ollama

QuantCoder transforms academic quant research into compilable QuantConnect LEAN algorithms using local LLMs. No cloud API keys required.

Models:

• qwen2.5-coder:14b — code generation, refinement, error fixing
• mistral — reasoning, summarization, chat

---

Installation

Prerequisites

• Python 3.10+
• Ollama running locally

# Pull the required models
ollama pull qwen2.5-coder:14b
ollama pull mistral

Setup

git clone https://github.com/SL-Mar/quantcoder-cli.git
cd quantcoder-cli

python -m venv .venv
source .venv/bin/activate

pip install -e .
python -m spacy download en_core_web_sm

Verify

# Check Ollama is running
curl http://localhost:11434/api/tags

# Launch QuantCoder
quantcoder

---

Usage

Interactive Mode

quantcoder        # or: qc

CLI Commands

# Search for papers
quantcoder search "momentum trading" --num 5

# Download and summarize
quantcoder download 1
quantcoder summarize 1

# Generate QuantConnect algorithm
quantcoder generate 1
quantcoder generate 1 --open-in-editor

# Validate and backtest (requires QC credentials)
quantcoder validate generated_code/algorithm_1.py
quantcoder backtest generated_code/algorithm_1.py --start 2022-01-01 --end 2024-01-01

Programmatic Mode

quantcoder --prompt "Find articles about mean reversion"

Autonomous Mode

quantcoder auto start --query "momentum trading" --max-iterations 50
quantcoder auto status

Evolution Mode (AlphaEvolve-inspired)

quantcoder evolve start 1 --gens 3 --variants 5
quantcoder evolve start 1 --gens 3 --push-to-qc   # Push best to QC
quantcoder evolve list
quantcoder evolve export abc123

Backtest with Detailed Metrics

# Shows Sharpe, Total Return, CAGR, Max Drawdown, Win Rate, Total Trades
quantcoder backtest generated_code/algorithm_1.py --start 2022-01-01 --end 2024-01-01

Library Builder

quantcoder library build --comprehensive --max-hours 24
quantcoder library status

---

Configuration

Configuration is stored in ~/.quantcoder/config.toml:

[model]
provider = "ollama"
model = "qwen2.5-coder:14b"
code_model = "qwen2.5-coder:14b"
reasoning_model = "mistral"
ollama_base_url = "http://localhost:11434"
ollama_timeout = 600
temperature = 0.5
max_tokens = 3000

[ui]
theme = "monokai"
editor = "zed"

QuantConnect Integration

For backtesting and deployment, set credentials in ~/.quantcoder/.env:

QUANTCONNECT_API_KEY=your_key
QUANTCONNECT_USER_ID=your_id

Remote Ollama

To use a remote Ollama instance:

[model]
ollama_base_url = "http://your-server:11434"

---

Architecture

quantcoder/
├── cli.py           # CLI entry point
├── config.py        # Configuration management
├── chat.py          # Interactive chat
├── llm/             # Ollama provider layer
├── core/            # LLM handler, processor, NLP
├── agents/          # Multi-agent system (Coordinator, Alpha, Risk, Universe)
├── evolver/         # AlphaEvolve-inspired evolution engine
├── autonomous/      # Self-improving pipeline
├── library/         # Batch strategy library builder
├── tools/           # Pluggable tool system
└── mcp/             # QuantConnect MCP integration

---

Background

QuantCoder was initiated in November 2023 based on "Dual Agent Chatbots and Expert Systems Design". The initial version coded a blended momentum/mean-reversion strategy from "Outperforming the Market (1000% in 10 years)", which received over 10,000 impressions on LinkedIn.

v2.0.0 is a complete rewrite — local-only inference, multi-agent architecture, evolution engine, and autonomous learning.

---

License

Apache License 2.0. See LICENSE.