MeMex

MeMex is designed to give Mac users a private visual memory, transforming their screen activity into an organized timeline of work moments. It is intended for individuals who struggle to recall specific tasks, applications, or information they have previously accessed on their computers. The core problem MeMex addresses is the fragmentation of modern work. Users often find themselves searching through scattered bookmarks, notes, screenshots, or browser tabs, wasting valuable time trying to locate past work. This leads to frustration and a loss of productivity when the exact context of a previous task is needed. MeMex captures screen activity and organizes it into a timeline, noting the time, applications used, and the nature of the activity. This allows users to return to specific moments without manual organization. A key feature is the ability to ask plain-English questions about past work, such as "What did I fix before shipping?", with answers directly linked to replayable screen evidence. Beyond recall, MeMex generates daily and weekly reports that summarize the work already completed, providing users with an overview of their activities. It also includes a feature called "Pattern Insights" which identifies repeated workflows. These insights can then be used to create reusable AI or agent playbooks, further enhancing automation and efficiency. Privacy is a central design principle. Raw recordings are stored locally on the user's Mac in a location they control. AI analysis is optional and can be configured to use a user-provided provider or a compatible local endpoint. Before any data is shared for third-party AI analysis, MeMex clearly explains what information will be sent and requests explicit user permission. MeMex operates by continuously recording screen activity. This activity is then processed into structured "timeline cards" containing titles, summaries, detailed observations, application/site context, and timestamps. These cards are stored locally in an SQLite database. When a user asks a question, MeMex parses the query and searches these local cards. Only the matched text and metadata are sent to the configured AI model to compose an answer, rather than sending raw screenshots or video for each query. The benefits for users include significantly reduced time spent searching for past work, improved recall of specific details and contexts, and a clearer understanding of their own work patterns. By automating the capture and organization of work history, MeMex frees up mental bandwidth and enhances overall productivity. Concrete use cases include quickly finding a specific file or window that was open earlier, recalling the exact steps taken to resolve a bug, or reviewing the progress made on a project over a specific period. Users can also leverage Pattern Insights to identify repetitive tasks that could be automated. MeMex is available for macOS as a one-time purchase of USD 19.99. The product focuses on privacy, with local storage and optional, user-controlled AI analysis. It does not currently list specific integrations or a detailed tech stack beyond its local SQLite database and optional AI provider configurations. In summary, MeMex acts as a private, searchable visual memory for your Mac, empowering users to effortlessly recall past work and identify opportunities for automation, all while maintaining strict control over their data.