Dayflow

What do they actually do

Dayflow is a native macOS app that passively captures your screen at 1 frame per second, stores short video chunks on your machine, and turns them into a human‑readable timeline you can review later. You install it via DMG/Homebrew/GitHub Releases, and all data is saved locally under ~/Library/Application Support/Dayflow/ with configurable retention rules (GitHub README).

Every 15 minutes, Dayflow batches recent frames and sends them to an AI engine you choose: a cloud model (e.g., Gemini with your API key), a CLI to ChatGPT/Claude, or a fully local LLM server like Ollama/LM Studio. The AI produces short activity “cards,” timelapses, distraction highlights, and you can export to Markdown; there’s also a beta “daily journal” for intentions and reflections. You can run the entire pipeline locally to avoid cloud traffic (GitHub README).

The project is open source (MIT) with visible community interest (thousands of GitHub stars and active releases) and the company is a YC Summer 2024 startup. Public discussion and roadmap notes are available on Hacker News and the YC page (GitHub README, Hacker News, YC company page).

Who are their target customer(s)

• Knowledge workers (engineers, PMs): They lose track of what happened across a busy day and find manual time tracking or note‑taking too interruptive; they want an automatic, reviewable timeline and simple exports for status updates (README/Roadmap).
• Freelancers and consultants who bill hourly: They spend time reconstructing billable work and answering client questions; they need lightweight evidence (timelines, timelapses, summaries) to streamline invoicing (README — Export/Timelapse).
• Privacy‑sensitive professionals (lawyers, therapists, executives): They want activity summaries without sending sensitive screenshots to third‑party servers; they require local storage and local model options (README — Data & Privacy).
• Managers and team leads: They need quick end‑of‑day summaries and trends (time on priorities) without repetitive report writing; they want auto timelines plus journaling and future dashboard tiles (README — Journal/Roadmap).
• Power users and open‑source enthusiasts: They dislike closed systems and want control over installation and model choice; they value MIT licensing, Homebrew/GitHub distribution, and local/cloud backend configurability (GitHub repo).

How would they acquire their first 10, 50, and 100 customers

• First 10: Personally invite top GitHub stargazers/contributors and active HN commenters to a private early‑adopter cohort, offering 30–60 minute setup sessions to validate install, privacy defaults, and the journal flow (GitHub README, HN launch).
• First 50: Publish a short how‑to video and config templates (cloud vs local), open issues/Discord, and announce onboarding slots to repo watchers, YC alumni Slack, and the HN thread to scale self‑serve setup and gather structured feedback (GitHub README, YC page).
• First 100: Do targeted webinars/demos for freelancers, consultants, and eng/product teams (exports, timelapse review, local model setup) and ask happy users for peer/team introductions; add a few influencer demos from privacy/power‑user accounts to lower trust friction.

What is the rough total addressable market

Top-down context:

Dayflow sits at the intersection of time tracking, personal “digital memory,” and local‑first productivity tooling. It overlaps with automatic trackers and screen‑recall tools but emphasizes image‑based summaries and local model options.

Bottom-up calculation:

Illustrative: assume 2 million macOS power users and privacy‑sensitive knowledge workers globally are open to automated day timelines; if 5% adopt at an average $10/month, that’s ~100,000 subscribers and ~$12M ARR potential. Broader use across freelancers and teams could expand this materially.

Assumptions:

• Focus on macOS users given today’s platform scope.
• Only a small fraction of the global knowledge‑worker base is power‑user/early‑adopter enough to try passive capture tools.
• Price point assumes a consumer/prosumer SaaS tier around $8–$12/month.

Who are some of their notable competitors

• Rewind: Continuously records screen/audio to build a searchable timeline and offers an “ask your history” AI layer; polished commercial product with desktop/mobile; raw recordings local but assistant can use cloud LLMs (Rewind site).
• RescueTime: Automatic activity tracker focused on app/website usage and productivity reports; strong for timesheets/analytics but does not capture screenshots/timelapse video (RescueTime features).
• Timing: Mac‑native automatic time tracker with interactive timeline, project assignment, and invoicing exports; relies on app/URL/document metadata rather than 1‑fps screen frames + VLM summaries (Timing features).
• ManicTime: Local‑first time tracker (Windows‑focused) that can capture periodic screenshots for playback and billing evidence; proprietary and lacks a built‑in image→AI summary pipeline (docs on screenshots).
• ActivityWatch: Open‑source, privacy‑first, cross‑platform tracker that records apps/window titles/tabs locally; closest OSS analogue for local data and extensibility but focused on structured activity rather than continuous screen captures or VLM summaries (ActivityWatch).