AgriPriceBD contains 1,779 daily retail mid‑prices for five Bangladesh commodities from July 2020 to June 2025, digitised with an LLM‑assisted pipeline. The authors benchmark seven forecasting methods, including SARIMA, Prophet, BiLSTM, and a Transformer. The dataset enables researchers and traders to compare classical and deep‑learning models for price prediction. Practitioners can now test models on realistic, high‑frequency data.