buffers.ai vs replenishment

buffers.ai is a web service for retail and manufacturing enterprises that automates demand forecasting, purchase planning, and inventory replenishment. Forecasting backtests TimesFM, LightGBM, Prophet, Exponential Smoothing, and Moving Average over a rolling six-month window, selecting a champion model for each SKU. Buy Plan creates a 12-month purchasing plan by category and month, in units and dollars, from a buying budget, including when sales history is absent. Replenishment recalculates each product’s target stock at every store daily and supports assortment, launches, promotions, and allocation across online, store, and wholesale channels. Production integrations use recurring batch feeds configured for a customer’s stack; CSV or XLSX uploads can support a benchmark without ERP work. The service uses HTTPS/TLS for the product and benchmark flow. Pricing starts at $25 a month for the 61+ store marginal plan, while other plans and possible setup fees vary by scope and scale.

replenishment is a Python library for simulating stock replenishment, optimizing policies, and visualizing policy behavior. It is designed for teams comparing policies against historical or synthetic demand. Listed optimization workflows include safety stock based on mean forecasts, k*RMSE or k*MAE, and percentile targets. For k*RMSE and k*MAE, safety stock can optionally increase when lead-time forecast quantities rise above a baseline. The library supports moving between CSV or DataFrame inputs and simulation configurations, and it can show visual comparisons of policies on replenishment time series. The README includes notebooks for policy variants, synthetic data generation, and loading stock replenishment data. Installation is through PyPI with `pip install replenishment`. The project is tested on Python 3.10, 3.11, 3.12, and 3.13; listed dependencies include pandas, Polars, PyArrow, matplotlib, ipykernel, and scienceplots. The package uses the MIT license and is classified as Alpha. The project welcomes contributions. Minimum order quantities, case-pack constraints, and promotion-aware demand handling are listed as areas for future contribution.