Vegetation photosynthesis, quantified as gross primary productivity (GPP), regulates the terrestrial carbon sink and land–atmosphere exchanges. At sub-daily scales, diurnal GPP dynamics reveal rapid adjustments to changing light, temperature and water conditions that are largely obscured in daily-to-annual aggregates, underscoring the need for developing global hourly GPP products. However, existing hourly products mostly rely on traditional machine-learning schemes that lack explicit biophysical constraints and an adequate representation of water limitation, leading to large uncertainties, especially in arid regions. Besides, the added value of hourly products for resolving diurnal behavior and responses to environmental stress remains poorly quantified. Here, we develop a causal knowledge-driven upscaling framework that couples the Peter and Clark Momentary Conditional Independence guided causal weights with ensemble learning strategies. Based on eddy-covariance measurements and multi-source meteorological variables, vegetation properties, and land-cover fields, we generated a global 0.05° hourly GPP product from 2000 to 2022, named EGO (Eddy covariance site-based Global hOurly) GPP, and then evaluated how well EGO reproduces observed diurnal cycles and their responses to extreme events. EGO GPP achieves an R² of 0.76 and an RMSE of 4.17 μmol CO₂ m⁻² s⁻¹ on independent test sites, and outperforms two recent hourly upscaling products (FLUXCOM and X-BASE; R² ≈ 0.60 and RMSE ≈ 5.5 μmol CO₂ m⁻² s⁻¹), with large improvement in drylands. EGO GPP clearly illustrates the diurnal progression of photosynthesis and captures observed diurnal metrics across diverse biomes, revealing strong midday depression and morning-skewed curves in drylands but near-symmetric cycles in high-latitude and humid tropical regions. Analyses of the June 2021 U.S. drought and the August 2003 European heatwave further show that EGO reliably tracks diurnal photosynthetic responses to extremes, including GPP reductions, earlier centroid/peak times and intensified midday depression, consistent with tower-based results. Looking ahead, EGO GPP provides a reliable foundation for investigating diurnal photosynthetic behavior, exploring vegetation–climate interactions and benchmarking Earth system models at a sub-daily scale. EGO GPP is available at <a href="https://doi.org/10.5281/zenodo.18253238" target="_blank" rel="noopener">https://doi.org/10.5281/zenodo.18253238</a> (Liu et al., 2026).