Browsing History

glebis/browsing-history

Query browsing history from all synced devices (iPhone, Mac, iPad, desktop). Supports natural language queries for filtering by date, device, domain, and keywords. Uses LLM classification for content categories. Can output to stdout or save as markdown/JSON to Obsidian vault.

8k tokens

context cost

the whole folder, loaded on every use

files

ships runnable scripts

copies elsewhere

how many repositories repackaged it

337

stars on the repo

on the repository, not the skill itself

Install

one command, takes just this skill from the repository

npx skills add https://github.com/glebis/claude-skills --skill browsing-history

What comes with it

26 066 bytes besides the instruction

.claude-plugin/plugin.json

README.md

browsing_query.py

scripts/init_db.py

scripts/sync_chrome_history.py

The instruction itself

14 sections, as written by the author

Browsing History Skill

Query browsing history from all synced devices with natural language.

When to Use

Use this skill when the user asks about:

Articles/pages they read (yesterday, last week, etc.)
Browsing history from specific devices (iPhone, iPad, desktop)
Finding pages by topic, domain, or keyword
Exporting browsing history to files
Grouping history by category or domain

Database

Location: ~/data/browsing.db

Synced devices: iPhone, iPad, Mac, desktop, Android

Timestamps

visit_time: Actual visit timestamp from Chrome (100% coverage for all devices)
first_seen: Import timestamp (fallback when visit_time unavailable)

The skill uses COALESCE(visit_time, first_seen) for accurate time-based queries.

Usage

python3 ~/.claude/skills/browsing-history/browsing_query.py "<query>" [options]

Options

| Option | Description | Example |

|--------|-------------|---------|

| --device | Filter by device | --device iPhone |

| --days | Number of days back | --days 7 |

| --domain | Filter by domain | --domain medium.com |

| --limit | Max results | --limit 50 |

| --format | Output format | --format json |

| --output | Save to file | --output history.md |

| --group-by | Group results | --group-by domain or --group-by category |

| --categorize | Use LLM to categorize | --categorize |

Example Queries

Basic queries:

# Yesterday's browsing history
python3 ~/.claude/skills/browsing-history/browsing_query.py "yesterday"

# Articles from iPhone yesterday
python3 ~/.claude/skills/browsing-history/browsing_query.py "yesterday" --device iPhone

# Last week's history grouped by domain
python3 ~/.claude/skills/browsing-history/browsing_query.py "last week" --group-by domain

# Find articles about economics
python3 ~/.claude/skills/browsing-history/browsing_query.py "economics" --days 7

Save to Obsidian:

# Save yesterday's history as markdown
python3 ~/.claude/skills/browsing-history/browsing_query.py "yesterday" \
  --output ~/Research/vault/browsing-2025-11-27.md

# Save with LLM categorization
python3 ~/.claude/skills/browsing-history/browsing_query.py "yesterday" \
  --categorize --group-by category \
  --output ~/Research/vault/browsing-categorized.md

# Save as JSON
python3 ~/.claude/skills/browsing-history/browsing_query.py "last week" \
  --format json --output ~/Research/vault/history.json

Device-specific:

# iPhone tabs
python3 ~/.claude/skills/browsing-history/browsing_query.py "yesterday" --device iPhone

# Desktop history
python3 ~/.claude/skills/browsing-history/browsing_query.py "today" --device desktop

# All mobile devices
python3 ~/.claude/skills/browsing-history/browsing_query.py "yesterday" --device mobile

Search and filter:

# Sites starting with "joy"
python3 ~/.claude/skills/browsing-history/browsing_query.py "joy" --days 7

# Medium.com articles
python3 ~/.claude/skills/browsing-history/browsing_query.py "last month" --domain medium.com

Natural Language Patterns

The script recognizes:

| Pattern | Interpretation |

|---------|----------------|

| yesterday | Previous day |

| today | Current day |

| last week | Past 7 days |

| last month | Past 30 days |

| last N days | Past N days |

Keywords are searched in URL and title.

Output Formats

Markdown (default)

# Browsing History: yesterday

*47 unique URLs from 2025-11-27*

## 2025-11-27

- [Article Title](https://example.com/article) - iPhone - 14:32
- [Another Page](https://another.com/page) - desktop - 16:45

Markdown with categories (--categorize --group-by category)

# Browsing History: yesterday

## News & Current Events
- [Breaking: Something Happened](https://news.com/...) - iPhone

## Technology & Programming
- [How to Build APIs](https://dev.to/...) - desktop

## Research & Learning
- [Academic Paper on AI](https://arxiv.org/...) - Mac

JSON (--format json)

{
  "query": "yesterday",
  "date_range": "2025-11-27",
  "total": 47,
  "results": [
    {"url": "...", "title": "...", "device": "iPhone", "time": "14:32", "category": "News"}
  ]
}

Workflow Examples

User: "Show me articles I read yesterday on my phone"

python3 ~/.claude/skills/browsing-history/browsing_query.py "yesterday" --device iPhone

User: "Save my browsing history from last week to Obsidian, grouped by category"

python3 ~/.claude/skills/browsing-history/browsing_query.py "last week" \
  --categorize --group-by category \
  --output ~/Research/vault/browsing-week.md

User: "Help me find that article about economics I read on my computer"

python3 ~/.claude/skills/browsing-history/browsing_query.py "economics" \
  --device desktop --days 7

User: "Sites that start with 'joy' from last week"

python3 ~/.claude/skills/browsing-history/browsing_query.py "joy" --days 7

Notes

URLs are deduplicated per day (same URL on same day = one entry)
visit_time: Actual visit timestamps from Chrome history
Desktop: 100% coverage (from Chrome SQLite last_visit_time)
Mobile: 100% coverage (extracted from Chrome Sync LevelDB)
first_seen: Fallback import timestamp (~15min resolution)
LLM categorization uses Claude 3.5 Haiku via llm CLI

How to use it

Copy the folder

Take glebis/browsing-history from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.