Most everyday experiences are multisensory, and all senses can trigger the conscious re-experience of unique personal events from the past embedded in their specific spatio-temporal context. Yet, little is known about how a cue’s sensory modality influences episodic memory retrieval, and which step of this process is impacted. This study investigated recognition and episodic memory across olfactory, auditory and visual sensory modalities in a laboratory-ecological task using a non-immersive virtual reality device presenting a three-room house where boxes delivered odours, musical pieces and pictures of face. At encoding, participants freely, actively and non-explicitly explored unique and rich episodes. At retrieval, participants were presented with modality-specific memory cues and were told to 1) recognize encoded cues among distractors and, 2) go to the room and select the box in which they encountered them at encoding. Memory performance and response times consistently revealed that music and faces outperformed odours in recognition memory, but that odours and faces outperformed music in evoking episodic context retrieval. Interestingly, correct recognition of music and faces was accompanied by more profound inspirations than was correct rejection of music and faces. By directly comparing memory performance across sensory modalities, our study provides further knowledge about specificities of olfactory memory and demonstrated that despite limited recognition, odours are powerful cues to evoke specific episodic memory retrieval.