Ziplime: Python algo trading kit
A new Python library, 'Ziplime', just launched for AI‑driven algo trading — it markets itself as a toolkit for building systematic strategies and backtests. (x.com)
PyPI shows Ziplime was published about four weeks ago and describes it as “Zipline reborn for 2026,” with native AI strategy generation and local, private execution. (pypi.org)) The project’s primary repositories are hosted on GitHub (Limex-com/ziplime and several forks), where the README repeats the pitch that AI can write strategy code and run backtests without copy‑paste glue code. (github.com)) The package replaces pandas and NumPy with Polars for its data layer, explicitly advertising dramatically faster data pipelines and improved performance on Apple Silicon. (pypi.org)) Ziplime advertises built‑in live trading support via a “Lime Trader” SDK so the same algorithm file can be used for backtesting and live execution. (pypi.org)) The project changelog lists version 1.13.19 changes including a pipeline refactor, a single execution clock to run handle_date once per simulation, an added vectorbt optional dependency group, and removal of a GRPC data source. (data.safetycli.com)) The package is available on PyPI for pip install and multiple community forks and alternate GitHub mirrors exist, indicating active development and community experimentation. (pypi.org))
