We consider QCD ttγ and ttZ production at hadron colliders as a tool to measure the ttγ and ttZ couplings. At the Tevatron it may be possible to perform a first, albeit not very precise, test of the ttγ vector and axial vector couplings in ttγ production, provided that more than 5 fb −1 of integrated luminosity are accumulated. The ttZ cross section at the Tevatron is too small to be observable. At the CERN Large Hadron Collider (LHC) it will be possible to probe the ttγ couplings at the few percent level, which approaches the precision which one hopes to achieve with a next-generation e + e − linear collider. The LHC's capability of associated QCD ttV (V = γ, Z) production has the added advantage that the ttγ and ttZ couplings are not entangled. For an integrated luminosity of 300 fb −1 , the ttZ vector (axial vector) coupling can be determined with an uncertainty of 45 − 85% (15 − 20%), whereas the dimension-five dipole form factors can be measured with a precision of 50 − 55%. The achievable limits improve typically by a factor of 2 − 3 for the luminosity-upgraded (3 ab −1 ) LHC.