Blogmon

Monitor developer blogs, extract insights, and build a knowledge base.

Installation

go install github.com/julienpequegnot/blogmon@latest

Or build from source:

git clone https://github.com/julienpequegnot/blogmon
cd blogmon
go build -tags sqlite_fts5 -o blogmon .

Note: The sqlite_fts5 build tag is required for full-text search support.

Quick Start

# Initialize configuration and database
blogmon init

# Add blogs to monitor
blogmon add https://jvns.ca
blogmon add https://fasterthanli.me
blogmon add https://brooker.co.za

# Fetch posts
blogmon fetch

# List posts
blogmon list

# Show post details
blogmon show 1

Commands

Command	Description
blogmon init	Initialize config and database
blogmon add <url>	Add a blog to monitor
blogmon fetch	Download new posts from feeds
blogmon extract	Extract insights from posts using LLM
blogmon score	Calculate community/relevance/novelty scores
blogmon link	Build concept graph by linking related posts
blogmon discover	Discover new blogs from references
blogmon trends	Show trending topics
blogmon list	List posts (--sort: date/score/source)
blogmon show <id>	Show post details
blogmon sources	List monitored sources
blogmon search <query>	Full-text search across posts
blogmon daemon	Run in daemon mode for auto-fetching
blogmon reindex	Rebuild full-text search index

Configuration

Config is stored in ~/.blogmon/config.yaml

interests:
  - topic: "distributed-systems"
    weight: 1.0
  - topic: "rust"
    weight: 0.9

scoring:
  community: 0.3
  relevance: 0.4
  novelty: 0.3

apis:
  llm_provider: "ollama"
  llm_model: "llama3.2"

fetch:
  concurrency: 5
  timeout_seconds: 30

daemon:
  interval_hours: 6

Architecture

Pipeline architecture:

fetch → extract → score → link → search
         ↑                        ↓
      daemon (scheduled)      query results

• fetch: Download posts from RSS feeds
• extract: Parse content, extract insights using LLM
• score: Calculate community/relevance/novelty scores
• link: Build concept graph linking related posts
• search: Full-text search with BM25 ranking

Development Status

Phase 1 (MVP) - Complete

• Project setup
• init command
• add command (with RSS feed auto-discovery)
• fetch command (with concurrent fetching)
• list command (with styled output)
• show command
• sources command

Phase 2 (Intelligence) - Complete

• LLM-powered insight extraction (Ollama)
• Community scoring (HN API)
• Relevance scoring (keyword matching)
• Novelty scoring (TF-IDF)

Phase 3 (Graph & Discovery) - Complete

• Concept graph (topic-based post linking)
• Blog discovery (from references)
• Trend detection (topic trending analysis)

Phase 4 (Polish) - Complete

• Full-text search (SQLite FTS5 with BM25 ranking)
• Daemon mode (scheduled background processing)
• Search command with ranked results
• Reindex command for index maintenance
• Enhanced list command with sorting