WorkLens — The Missing Layer for Knowledge Workers

RECUR-18348. I still remember the ticket number. It was a bug that took me three hours to fix — not because the fix was hard, but because I spent two of those hours just figuring out what state the system was in when the bug was introduced, who touched that code last, what decision we'd made three weeks ago that created this situation. That ticket broke something in me. We were losing hours every day not to hard problems, but to the sheer friction of reconstructing context. I started building WorkLens the same week. ## The context-switching tax Every knowledge worker pays this tax and nobody talks about it. You're deep in a problem, get pulled into a meeting, come back, and spend 20 minutes just getting back to where you were. Multiply that by 4-5 times a day. Multiply by 5 days a week. That's hours of pure waste — not procrastination, not distraction, just the mechanical overhead of context reconstruction. Tools like Jira, Linear, Notion — they're for tracking work status, not work context. They tell you what tickets are open, not why decisions were made, what was tried, what dead ends were hit. The actual work context lives in your head, in Slack threads nobody can find, in Zoom calls nobody recorded, in code reviews that are closed and forgotten. And this isn't just a developer problem. A designer coming back to a project they haven't touched in two weeks. A product manager trying to remember why a feature was scoped the way it was. A finance analyst picking up a model they built six months ago. Everyone suffers from this. Yet the tooling industry has basically ignored it. ## What WorkLens actually does WorkLens is a passive work context capture system. You install it, and it starts watching. File activity, folder changes, which apps you're switching between, how long you spend on things. It builds a timeline of your actual work — not what you logged, but what you did. The real-time dashboard shows you your work timeline, groups activity into sessions automatically, and surfaces patterns. When did you last touch this file? What other files were you looking at when you made that change? Who else worked on this component, and when? The memory graph is where it gets interesting. WorkLens builds a graph of relationships — between files, between sessions, between decisions and outcomes. When you come back to a project, it can surface: "Last time you worked on this, you were also editing these config files. You left a note here. The build was failing because of this dependency change." ### Passive capture is the key This is the design decision I'm most committed to. No logging prompts. No "write a daily standup." No workflow changes. WorkLens captures passively or it fails. The moment you ask people to do active input, you've already lost — people have tried that (journals, time trackers, daily notes apps) and 90% of people abandon them within a week. The file/folder watcher runs in the background. It's low-overhead. The AI layer processes the captured data asynchronously — it's not blocking your work to build context about your work. Proactive suggestions surface in a side panel that you can ignore until you need them. ## Why this gap exists I've been thinking about why no one has built this at scale. Partly it's a privacy problem — capturing work context means capturing everything, and that's scary. Partly it's because the companies that could build this (Microsoft, Google, Apple — they already have access to all this data) have no incentive to. They'd rather you stay in their ecosystem, not get smarter about leaving it. The developer tools space has been getting better — GitHub Copilot, Cursor, linear's automations. But those tools are built for devs. The average knowledge worker has nothing. A project manager, a recruiter, an operations analyst — their AI tooling is basically "here's ChatGPT, figure it out." WorkLens is for all of them. ## What I'm building toward The current version is the foundation — passive capture, timeline, real-time dashboard. What I'm building toward is a proactive intelligence layer. Not just "here's what you did" but "here's what you should know right now." Surfacing a decision from three months ago that's relevant to what you're doing today. Flagging when a pattern of work suggests you're about to hit a problem you hit before. Connecting dots across your work history that you'd never connect manually. Think of it as an AI that watches you work not to judge you, but to become the institutional memory that every team loses when people leave. The knowledge that lives only in senior engineers' heads. The decisions that were made in a Slack DM in 2022 that nobody can find. WorkLens captures that — if you deploy it before it's too late. It's early. The repo is open, the foundation is solid. I'm building this in the open because I want feedback from people who feel this pain the way I felt it staring at RECUR-18348. GitHub: https://github.com/fl-satyakam/worklens