GRB 201015A is a peculiarly low luminosity, spectrally soft gamma-ray burst (GRB), with T90 = 9.8 ± 3.5 s (time interval of detection of 90 per cent of photons from the GRB), and an associated supernova (likely to be type Ic or Ic-BL). GRB 201015A has an isotropic energy $E_{\gamma , \rm iso}$$= 1.75 ^{+0.60} _{-0.53} \times 10^{50}$ erg, and photon index $\Gamma = 3.00 ^{+0.50} _{-0.42}$ (15-150 keV). It follows the Amati relation, a correlation between $E_{\gamma , \rm iso}$ and spectral peak energy Ep followed by long GRBs. It appears exceptionally soft based on Γ, the hardness ratio of HR = 0.47 ± 0.24, and low-Ep, so we have compared it to other GRBs sharing these properties. These events can be explained by shock breakout, poorly collimated jets, and off-axis viewing. Follow-up observations of the afterglow taken in the X-ray, optical, and radio reveal a surprisingly late flattening in the X-ray from t = (2.61 ± 1.27) × 104 s to $t = 1.67 ^{+1.14} _{-0.65} \times 10^6$ s. We fit the data to closure relations describing the synchrotron emission, finding the electron spectral index to be $p = 2.42 ^{+0.44} _{-0.30}$ and evidence of late-time energy injection with coefficient $q = 0.24 ^{+0.24} _{-0.18}$. The jet half opening angle lower limit (θj ≥ 16°) is inferred from the non-detection of a jet break. The launch of SVOM and Einstein Probe in 2023 should enable detection of more low-luminosity events like this, providing a fuller picture of the variety of GRBs.