Dayflow Export
https://github.com/obra/dayflow-export- Category
- Productivity
- Rank
- No. 2075Tools index
Previous survey · No. 2070 ·
- Pricing
- Open Source
- Type
- TOOL
- Builder
- obra
- GitHub
- 12 stars
- Date
About
Export Dayflow activity data to markdown transcripts for LLM consumption.
What it does
It reads Dayflow’s local SQLite records and builds one readable activity journal per day. Each journal combines timeline entries, categories, apps or sites, summaries, timestamped observations, sampled screenshot references, and recorded distractions.
Why it's ranked here
This is a focused, transparent utility with little setup burden. It handles incremental exports, single days, and date ranges while preserving useful context beyond summaries. Its value depends on already having Dayflow’s database and recordings.
What's good
The tool uses only Python’s standard library, so installation stays simple. It validates required data, ignores deleted timeline cards and screenshots, samples images evenly, checks that linked recordings exist, and tolerates malformed metadata by falling back safely.
Tradeoffs
Installation means cloning the repository rather than using a package manager. It requires both the Dayflow SQLite database and recordings directory. Screenshots remain relative links instead of embedded copies, and each observation period includes at most five sampled images.
How to use it well
It suits Dayflow users who want dated, portable activity journals for later reading or language-model prompts. Run incremental exports routinely, then target specific dates or ranges when rebuilding history. It does not capture activity, replace Dayflow, or perform downstream analysis itself.
Technical notes+
The dayflow-export Python executable uses argparse, sqlite3, json, re, datetime, and pathlib. It queries timeline_cards, observations, screenshots, and batch_screenshots from chunks.sqlite, filters deleted records, derives an observation timestamp offset from each card’s displayed start time, and writes daily Markdown. sample_screenshots() selects up to five evenly spaced records per observation. parse_metadata() converts invalid JSON to an empty dictionary. README.md documents incremental, single-date, and date-range CLI modes plus configurable output and recording-link paths.
Observed
- License
- MIT
- Primary language
- Python 3
- Dependencies
- Python 3 standard library only
- Install surface
- Clone the Git repository; no package-manager installation is documented
- Interface
- Command-line executable
- Input
- Dayflow SQLite database and recordings directory
- Output
- Daily Markdown files with relative screenshot links
- Platform documentation
- A typical macOS Dayflow data location is documented
Read from README.md, dayflow-export.
What it can do
Export Dayflow activity data to markdown format
Dayflow activity data → Markdown transcripts
Convert activity logs for LLM processing
Dayflow activity logs → LLM-compatible data format
Generate structured transcripts from time tracking data
Time tracking activities from Dayflow → Structured markdown documents
Transform personal productivity data into readable format
Personal activity records → Human-readable markdown text
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.