new file mode 100644
index 0000000..af2d5ef
@@ -0,0 +1,30 @@
+---
+type: idea
+title: brain rewinder
+tags: [memory, hardware, wearable, bci]
+status: explored
+first_captured: 2025-12-22
+updated: 2026-04-10
+sources:
+ - sources/ideaflow/2025-12-22_brain-rewinder-to-remember-previous-thoughts-probably-str.md
+ - sources/apple-notes/archived/potential-projects-archived.md
+ - sources/google-sheets-ideas.md
+---
+
+# brain rewinder
+
+a device or system that helps you remember previous thoughts — the experience of losing a thought mid-stream and wanting to rewind your brain to retrieve it. the core insight is that existing solutions like Limitless and Plaud Note are device-based and record externally, but the ideal solution would either be a very cheap/easy device or a non-device approach entirely.
+
+one route explored is sensory triggering — using smells, sounds, or visuals to invoke prior mental states. smell-based recall has some research backing (Harvard Mind & Mood), though there is a risk of false recall ("might be cooked if you hallucinate the thing and ur sad"). AlterEgo from MIT was also considered as a silent speech interface that might bridge the gap.
+
+**spreadsheet evaluation:** originality 9/10, excitement 10/10 (highest tier). competitive landscape: academic research exists on olfactory memory reactivation (ACM papers, wearable prototypes), Rewind.ai records screen/audio but is visual-log-based, no consumer product systematically uses sensory triggers for thought retrieval. MVP estimated at 4-8 weeks — could start as a simple app associating scents/sounds with contexts and replaying them as recall triggers, though hardware adds complexity. tech depth ranked 6/10, labeled [WRONG FIT] for a 2-month build ("cool concept but no clear technical path in 2 months — what would you even build?"). blue ocean analysis: BLUE OCEAN positioning with medium 2-month feasibility.
+
+this connects to the broader [[episodic-memory-builder|episodic memory builder]] idea and to [[sensor-capturer|sensory capture]] more generally. the [[universal-data-capturer|universal data capturer]] could also serve as infrastructure for this — logging enough context that an AI could reconstruct what you were thinking about. the [[always-on-ai-assistant|always-on AI assistant]] approach (always listening, always transcribing) may be the most practical path.
+
+---
+
+## timeline
+
+- [2025-12-22] initial capture in ideaflow — explored device vs non-device approaches, sensory routes, existing products
+- [2025-12-22] also captured in apple notes potential-projects-archived (strikethrough, moved to ideaflow)
+- [2026-04-10] google sheets evaluation — 9/10 originality, 10/10 excitement, BLUE OCEAN, [WRONG FIT] for 2mo scope