Skip to content

bm - Bookmark Manager

CLI tool for creating a searchable database of browser bookmarks with both text (FTS5) and semantic (Ollama embeddings) search.

Features

  • Chrome Import - Auto-discovers all Chrome profiles, imports bookmarks with stable identity tracking
  • Full-Text Search - SQLite FTS5 for instant text search across titles, URLs, and page content
  • Semantic Search - Find bookmarks by meaning using local Ollama embeddings
  • Page Fetching (beta, optional) - Downloads and extracts text content from bookmarked pages for deeper search
  • Statistics - Breakdown by year, fetch status, and profile
  • Interactive Mode - TUI browser with search, sorting, and filtering

Installation

bm

go install github.com/GiGurra/bm@latest

Semantic search requires Ollama running locally with an embedding model.

# Install Ollama
brew install ollama          # macOS / Linux (Homebrew)
# or download from https://ollama.com/download

# Start the Ollama server
ollama serve                 # or: brew services start ollama

# Pull the default embedding model
ollama pull qwen3-embedding:0.6b

Text search (bm search) and the interactive browser (bm list -w) work without Ollama. Only semantic search (bm search -s, bm index) requires it.

Quick Start

# Import bookmarks and build search index
bm sync                                  # or: bm import && bm index

# Search
bm search "golang tutorials"
bm search -s "how to learn programming"  # semantic search

# Optional: also fetch page content for deeper search (beta)
bm sync --fetch

# Interactive browser
bm list -w

# Stats
bm stats

How It Works

Chrome Bookmarks JSON
        │
        ▼
   bm import ──► SQLite DB (~/.bm/bm.sqlite)
        │                │
        │          (optional) bm fetch ──► page content stored in DB
        │                │
        └───► bm index ──► Ollama embeddings stored in DB
                         │
               bm search / bm list -w
  1. Import reads Chrome's bookmark files and stores them in SQLite with a composite key (url, folder_path, source) — duplicate URLs in different folders are preserved
  2. Index generates vector embeddings via Ollama for semantic search (works on titles/URLs/folders alone)
  3. Fetch (optional, beta) downloads page content and extracts text from HTML for deeper search
  4. Search queries either FTS5 (text) or cosine similarity (semantic)

Data Storage

All data is stored locally in ~/.bm/bm.sqlite using WAL mode for performance. The database uses a pure-Go SQLite driver (modernc.org/sqlite) — no CGO required.