Practical overview
When it fits
Use it when you need an inspectable Python workflow that separates offline model training from online prediction and supports configurable signal destinations.
Who it is for
- Python developers building or adapting machine-learning trading workflows.
- Users experimenting with cryptocurrency signals, feature engineering and historical backtesting.
- Teams connecting model predictions to notifications, APIs, databases or transaction execution.
Documented uses
- Download and combine historical market data.
- Generate custom or indicator-based features.
- Train models and produce predictions and signals.
- Backtest trading performance on historical data.
- Send online outputs or execute configured transactions.
Requirements & limits
The repository is public and specifies an MIT license. The documented signal service publishes example signals through Telegram.
- Python is the primary language and the project is described as a CLI-capable application.
- Offline training must be completed and model files made available before the online signaler can work.
- Feature definitions, symbols and frequencies are controlled through configuration.
- The documented example uses Binance, BTCUSDT and one-minute analysis.
- Transaction execution is available through customizable output functions, but the evidence does not establish a particular deployment result or trading performance.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Prepare models and signals from historical data
Run the documented offline pipeline to create model outputs and trading signals from configured data.
Before you start
- Python environment for the repository scripts.
- A configuration file, such as one of the samples in the config folder.
- Configured data sources and a data folder for offline batch files.
Steps
Run the download script with the configuration file.
Run merge to align downloaded records by timestamp into one table.
Run features and labels to create derived inputs and training targets.
Run train, predict, signals and output with the same configuration file.
Expected outcome
The documented batch sequence produces trained model files, predictions and signal outputs for use by the service.
Guide limitations
- The project documentation says model files must be available before the signaler service can work.
- The documented data sources are Binance and Yahoo.
- This recipe describes the repository's batch sequence, not a verified live-trading result.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 30 September 2026
- Original website title
- asavinov/intelligent-trading-bot · Official homepage title · retained in its source language