TradingView Strategy Optimizer

A Chrome extension and local Python backend that optimizes TradingView strategy parameters using Optuna-powered Bayesian optimization. The extension runs as a side panel on TradingView chart pages, communicates with a local FastAPI backend over WebSocket, and iteratively tunes your strategy inputs to find optimal parameter combinations.

Prerequisites

• Python ≥ 3.12, < 3.14
• Node.js ≥ 20.19.0, < 25
• uv — fast Python package/project manager
• npm ≥ 10
• PowerShell 7+ (pwsh) — used by the helper scripts

Quick start

Start both the backend and the extension dev server in one command:

pwsh -File scripts/dev.ps1

Then load the extension in Chrome:

1. Go to chrome://extensions and enable Developer mode.
2. Click Load unpacked and select the extension/dist folder.
3. Open a TradingView chart and click the extension icon to open the side panel.

Manual setup

Backend

# Install dependencies
uv sync --frozen --project backend

# Run the server
uv run --project backend uvicorn backend.main:app --host 127.0.0.1 --port 8765 --reload

Extension

cd extension
npm ci
npm run dev

Optional backend extras

# AutoSampler (installs optunahub, scipy, torch, and cmaes)
uv sync --frozen --project backend --extra auto-sampler

# Gaussian Process sampler (installs scipy and torch)
uv sync --frozen --project backend --extra sampler-gp

# CMA-ES sampler (installs cmaes)
uv sync --frozen --project backend --extra sampler-cmaes

If auto-sampler is not installed, sampler=auto falls back to deterministic TPE.

Scripts

Script	Purpose
pwsh -File scripts/dev.ps1	Launch backend + extension dev server
pwsh -File scripts/check.ps1	Run all linting, type-checking, tests, and audits for both backend and extension

Extension commands

cd extension
npm run dev             # Vite dev server with HMR
npm run build           # Production build
npm run typecheck       # TypeScript type-check
npm run lint            # ESLint (zero warnings)
npm run lint:fix        # ESLint with auto-fix
npm run test            # Vitest tests
npm run format          # Prettier format
npm run format:check    # Prettier check
npm run ci              # Full CI suite (typecheck + lint + test + audit)

Backend commands

# Linting
uv run --project backend --extra dev ruff check --config backend/pyproject.toml backend

# Type-checking
uv run --project backend --extra dev pyright -p backend/pyrightconfig.json backend

# Tests
uv run --project backend --extra dev python -m pytest -q backend/tests

Project structure

├── backend/                Python backend (FastAPI + Optuna)
│   ├── main.py             WebSocket server & API endpoints
│   ├── optimizer.py        Optuna ask-and-tell optimizer wrapper
│   ├── models.py           Pydantic message schemas
│   ├── data/               SQLite study storage (git-ignored)
│   ├── tests/              Backend tests
│   └── pyproject.toml      Python dependencies & config
├── extension/              Chrome extension (React + Vite)
│   ├── src/
│   │   ├── manifest.json   MV3 extension manifest
│   │   ├── background/     Service worker
│   │   ├── content/        Content script (TradingView pages)
│   │   ├── sidepanel/      React side panel UI
│   │   ├── shared/         Shared types/utilities
│   │   └── utils/          Utility modules
│   ├── package.json        Node dependencies & scripts
│   └── vite.config.ts      Vite + CRXJS config
└── scripts/                PowerShell helper scripts

How it works

1. The content script injects into TradingView chart pages and interacts with the strategy tester UI.
2. The side panel provides the optimizer controls and connects to the backend via WebSocket (ws://127.0.0.1:8765).
3. The backend manages Optuna studies with an ask-and-tell loop — it suggests parameter combinations, receives performance results, and converges toward optimal values.
4. Studies are persisted locally as SQLite databases in backend/data/.

Supported samplers

Sampler	Extra required	Description
tpe	—	Tree-structured Parzen Estimator (default)
random	—	Random sampling
qmc	—	Quasi-Monte Carlo
auto	auto-sampler	OptunaHub AutoSampler (falls back to TPE if not installed)
gp	sampler-gp	Gaussian Process (installs scipy and torch)
cmaes	sampler-cmaes	CMA-ES evolutionary strategy (installs cmaes)

License

This project is licensed under the GNU General Public License v3.0.

---

Disclaimer: This software is provided for research and educational purposes only. The author is not responsible for financial losses resulting from the use of this software.