Unlost

Unlost enters the productivity tool landscape with a deceptively simple promise: remember everything you see and hear on your computer, and make it searchable with natural language. It is not a screen recorder in the conventional sense, nor is it a note-taking app. Instead, Unlost operates as a continuous, passive memory layer that indexes your digital activity across applications, from web browsing to Zoom calls, and then surfaces specific moments through a search interface. For professionals who frequently lose track of details scattered across tabs, meetings, and documents, this tool offers a compelling shortcut to recall. However, its value hinges on whether the convenience of total recall justifies the privacy and performance trade-offs inherent in always-on recording.

Unlost’s strongest differentiator is its local-first architecture. Data collection, processing, and storage occur entirely on the user’s machine, with the company explicitly stating they only collect an email address. This positions Unlost as a privacy-conscious alternative to cloud-based memory tools, which typically require uploading sensitive screen captures or audio. For researchers handling confidential data, consultants managing client information, or students concerned about surveillance, this local approach is a significant advantage. The trade-off is that the tool currently supports only Mac M1 and M2 devices, and details about storage consumption and CPU impact remain undisclosed. Heavy users with limited disk space or older hardware may find these unknowns problematic.

In practice, Unlost’s core workflow is straightforward: the tool runs in the background, capturing screen content and audio, then uses AI to extract text, recognize images, and index conversations. Users can later search using natural language queries like “what was the Q3 budget figure from the client meeting?” or filter by app or time range. The inclusion of keyboard shortcuts aims to make retrieval frictionless, allowing users to summon search without leaving their current task. Meeting transcript support for Zoom, Google Meet, and Teams is a standout feature, as it transforms ephemeral conversations into searchable text. However, the accuracy of multi-speaker transcription and the ability to handle accents or technical jargon remain open questions that would need rigorous testing.

Unlost is best suited for individuals who consume and generate large volumes of digital information and need to retrieve specifics quickly. Researchers can use it to pinpoint a data point from a week’s worth of browsing; students can search lecture recordings for a key concept; project managers can recall action items from a series of cross-app meetings. Yet, the tool’s value diminishes for users who primarily work in a single app or have disciplined note-taking habits. The lack of cloud sync means that switching computers or working across multiple devices is not supported, which could be a dealbreaker for team environments or users with both a desktop and laptop.

Potential adopters should weigh the convenience against the uncertainties. The absence of clear storage and energy consumption metrics makes it difficult to assess long-term impact on system performance. Privacy controls, including app exclusion, are present but their granularity is not fully detailed—can a user prevent sensitive apps like banking or personal messaging from being recorded? The answer to that question will determine whether Unlost is a trustworthy assistant or an overreaching observer. For now, Unlost is a promising but niche tool: a local memory prosthetic for power users on modern Macs who prioritize privacy and are willing to accept platform limitations and undisclosed resource usage.