We have undertaken a meta-analysis of the English literature, to assess the component alignment outcomes after imageless computer assisted (CAOS) total knee arthroplasty (TKA) versus conventional TKA. We reviewed 23 publications that met the inclusion criteria. Results were summarised via a Bayesian hierarchical random effects meta-analysis model. Separate analyses were conducted for prospective randomised trials alone, as well as for all randomised and observational studies. In 20 papers (4,199 TKAs) we found a reduction in outliers rate of approximately 80% in limb mechanical axis when operated with the CAOS. For the coronal femoral and tibial implants positions, the analysis included 3,058 TKAs. The analysis for the femoral implant showed a reduction in outliers rate of approximately 87% and for the tibial implant a reduction in outliers rate of approximately 80%. Imageless navigation when performing TKA improves component orientation and postoperative limb alignment. The clinical significance of these findings though has to be proven in the future.