Dayflow

Product Introduction

1. Definition: Dayflow is a local-first, AI-powered automatic work journal and productivity analytics application for macOS. It operates as a passive screen recorder and contextual activity analyzer.
2. Core Value Proposition: It exists to solve the problem of "work amnesia" for knowledge workers by automatically documenting daily accomplishments. Its core value is providing an auditable, private, and automated record of work for performance reviews, standup updates, and personal productivity analysis without manual time tracking.

Main Features

1. Automatic, Context-Aware Activity Logging: Dayflow records the screen at 1 frame per second and uses configured AI models to analyze activity in 15-minute intervals. Unlike simple app trackers, it uses computer vision and LLMs to understand context—differentiating between "researching on YouTube" and "watching entertainment."
2. Local-First & Open Source Architecture: The application is built on a principle of user sovereignty. All raw data (screenshots, databases) is stored locally on the user's Mac. The code is open source under the MIT license, allowing full auditability, customization, and self-hosting.
3. Flexible AI Model Integration: Users can configure any AI provider for analysis. This includes local models (via Ollama or LM Studio) for maximum privacy, or cloud models (OpenAI's ChatGPT, Anthropic's Claude, Google's Gemini) for higher accuracy. The system is provider-agnostic.
4. Multi-Fidelity Work Review Interface: The app provides several views to review captured work: a visual Timeline of the day, a text-based Daily Summary, a Weekly Summary for broader progress tracking, and a Chat interface for querying the journal with natural language.
5. One-Click Update Generation: Leveraging the AI-generated summaries, Dayflow can instantly produce ready-to-share updates for standup meetings or one-on-ones, transforming raw activity data into narrative progress reports.

Problems Solved

1. Pain Point: The inability to accurately recall and quantify work completed over weeks or months, leading to weak self-promotion materials during performance reviews and promotion cycles.
2. Target Audience: Primary personas include software engineers, developers, freelance knowledge workers, product managers, and researchers—any professional whose output is digital and who needs to report on abstract, project-based work.
3. Use Cases: Essential for preparing biannual performance review packets; generating daily standup updates without relying on memory; conducting personal weekly reviews to audit time allocation; and maintaining a private, searchable archive of all work-related problem-solving and decisions.

Unique Advantages

1. Differentiation: Versus traditional time trackers (e.g., Toggl, RescueTime), Dayflow provides contextual understanding, not just app/website logging. Versus manual journaling apps, it is fully automatic and evidence-based. Its open-source, local-first model contrasts with purely SaaS-based competitors.
2. Key Innovation: The combination of lightweight, continuous screen capture with configurable AI analysis to produce a searchable, contextual narrative of a workday. The technical innovation is its modular AI pipeline that can run entirely on-device, ensuring privacy without sacrificing core functionality.

Frequently Asked Questions (FAQ)

1. Is Dayflow a keylogger or screen monitoring spyware? No. Dayflow is a local-first application where the user retains full control. All data processing can be configured to occur entirely on your local machine using local AI models. The code is open-source for public audit, and recordings are never transmitted to Dayflow's servers.
2. How does Dayflow's AI understand different work contexts on the same application? It uses a combination of screen capture (computer vision) and Large Language Model (LLM) analysis. The AI examines visual content, text snippets, and application states to infer context—for example, distinguishing a coding session in VS Code from reading documentation in a browser, even if both are on the same monitor.
3. What are the system requirements for running Dayflow with local AI? Dayflow requires macOS 13.0 (Ventura) or later. For the core app, resource use is minimal (~100MB RAM). However, running local LLMs (like via Ollama) requires sufficient GPU memory (VRAM) and is more demanding on battery; it is recommended for use while plugged in during intensive sessions.
4. Can I export or backup my Dayflow journal data? Yes. Since all data—including SQLite databases and screenshot archives—is stored locally in your macOS application support folder, you can use standard system backup tools (Time Machine, cloud sync of the folder) to secure your journal. The open-source nature also means you can build custom export tools.
5. Is Dayflow free to use, and what are the costs for AI? The Dayflow application itself is free and open-source. Costs depend on your AI configuration: using local models (Ollama) is free; using Google's Gemini may fall under its free tier; using OpenAI's ChatGPT or Anthropic's Claude API requires a paid subscription with those providers to access their APIs.