Lost dream memories — AI + sleep-sensor capture to recall & log

Many people who are interested in creative work, emotional processing, or sleep health routinely lose their dream memories; an estimated 200 million global users could be receptive to tools that improve recall and logging. Current approaches—manual journaling and post-wake interviews—suffer from poor compliance and recall bias, leaving therapists, researchers and self-trackers with sparse, noisy data. A practical product would combine affordable EEG (headband or ear-EEG) plus high-quality wrist sensors and nocturnal audio capture with multimodal AI that maps latent sleep signals and short user reports into coherent, timestamped dream reconstructions and a searchable dream log. Delivered as a $5/month ($60/yr) subscription with opt-in privacy controls, features could include automated micro-awaken prompts, morning synthesis narratives, trend analytics, and optional clinical export for therapists. The market is attractive now: total addressable market ~$12.0B (200M users × $60/yr), with a Market Score of 92/100 and Revenue Potential 88/100, driven by falling sensor costs and maturing AI language and multimodal models that materially improve mapping between physiologic signals and reported content. Consumer willingness to pay for personalized sleep and mental wellness subscriptions is increasing, lowering go-to-market friction compared with prior years. To stand out you would need to prioritize signal quality, validated accuracy, transparent privacy-by-design, and partnerships with sleep clinics and research labs to deliver credible, evidence-backed features rather than speculative reconstructions. Key challenges are inherent—dream content is subjective and unverifiable, models risk hallucination, regulatory and ethical concerns about sensitive mental content are real, and consumer retention will hinge on demonstrable utility beyond novelty.