Brain-computer interface (BCI) systems allow users to interact with the environment by decoding their brain signals with applications in various fields, such as alternative communication. A promising control signal for BCI systems is code-modulated visual evoked potentials (c-VEPs), which have shown improved accuracy and faster selection times compared to other control signals. The integration of mixed reality (MR) with BCI systems has the potential to enhance usability and accessibility by combining real and virtual elements in real time. However, no previous studies have explored this combination with c-VEPs. This pilot study aims to evaluate the integration of MR with c-VEP-based BCI spellers and analyze visual fatigue compared to conventional screens. Twelve users tested a 36-character speller in both MR and a traditional screen condition. Accuracy, information transfer rate (ITR), usability, and visual fatigue were assessed. Results showed 99.39% accuracy in MR and 95.32% on conventional screens, with ITRs of 28.95 bpm and 26.83 bpm, respectively. No significant differences in visual fatigue were observed, which was not high in either the MR or the traditional screen condition, and usability received an average system usability scale (SUS) score of 86.90. In conclusion, integrating MR with c-VEP-based BCI systems is feasible, achieving performance levels comparable to conventional setups, high usability, and no significant increase in visual fatigue. These results highlight the promising potential of this approach for practical applications.