RSS Summarizer

AI-powered RSS feed aggregator that summarizes articles using Poolside Laguna models (recommended) or Claude, rates them by importance, and helps you stay informed without the noise.

Features

• AI Summarization: Auto-summarize articles with Poolside Laguna models (recommended) or Anthropic Claude
• Importance Scoring: Articles rated 1-5 based on your interests
• Smart Organization: Save for later, archive, mark as read
• Topic Detection: Auto-categorize with normalized broad topics
• Intelligent Filtering: By importance, topic, feed, read/saved/archived status
• Background Processing: Temporal workflows poll feeds and process articles automatically

Quick Start

Prerequisites

• Docker & Docker Compose
• Go 1.26+ and Node.js 20+ (for local development)
• Poolside API key (get one here) - Recommended for best performance
• (Alternative) Anthropic API key (get one here)

Setup

# 1. Clone repo
git clone <repo-url>
cd rss-summarizer

# 2. Configure backend
cp backend/.env.example backend/.env
# Edit backend/.env and add your API key:
# Option A (Recommended): Poolside API key (better performance)
# LLM_API_KEY=ps-your-key-here
# LLM_MODEL=poolside/laguna-m.1
# Provider: Set to "poolside" in admin UI

# Option B: Anthropic API key
# LLM_API_KEY=sk-ant-api03-your-key-here
# LLM_MODEL=claude-3-5-sonnet-20241022

# 3. Start everything
make start

That's it! Open http://localhost:5173

The make start command:

• Starts PostgreSQL + Temporal (Docker)
• Runs database migrations
• Starts backend API (port 8080)
• Starts frontend dev server (port 5173)

Development

# Daily workflow
make start          # Start everything
make stop           # Stop all services

# Run components separately
make dev-backend    # Backend only (requires Docker running)
make dev-frontend   # Frontend only (requires backend)

# Testing
make test           # Run all tests

# Database
make migrate-up     # Apply migrations
make migrate-down   # Rollback last migration

# Other
make help           # Show all commands

Tech Stack

Backend: Go, PostgreSQL, Temporal, Huma v2 (OpenAPI auto-generation) Frontend: SvelteKit, Svelte 5, TypeScript, Tailwind CSS AI: Poolside Laguna models (recommended) or Anthropic Claude API

Project Structure

rss-summarizer/
├── backend/
│   ├── cmd/api/           # Entry point
│   ├── internal/
│   │   ├── api/           # HTTP handlers (Huma)
│   │   ├── domain/        # Business models
│   │   ├── repository/    # Data access (Postgres)
│   │   ├── service/       # LLM, RSS, scraper
│   │   └── workflow/      # Temporal workflows
│   ├── migrations/        # SQL migrations
│   └── CLAUDE.md          # Backend dev guide
├── frontend/
│   ├── src/
│   │   ├── lib/           # Components, stores, API client
│   │   └── routes/        # SvelteKit pages
│   └── CLAUDE.md          # Frontend dev guide
├── prd.md                 # Product requirements
└── Makefile               # Unified commands

Key Concepts

Articles & User State

• Articles are global (shared across users)
• User state is per-user:
  • is_read - Track what you've read
  • is_saved - Save for later reading
  • is_archived - Archive to remove from main view

Topics

• Auto-detected by LLM and normalized to broad categories (e.g., "Go", "Security", "AI")
• Set preferences per topic: "High Interest", "Normal", or "Hide"
• Topics are global, preferences are per-user

API Client

• Frontend TypeScript client auto-generated from backend OpenAPI spec
• Backend changes → Regenerate client: npm run generate:api (from frontend dir)
• See frontend/CLAUDE.md for details

Configuration

Required Environment Variables

Backend (backend/.env):

DATABASE_URL=postgres://rss_user:rss_pass@localhost:5432/rss_summarizer?sslmode=disable

# Option 1 (Recommended): Poolside API key
# Get from https://platform.poolside.ai/
LLM_API_KEY=ps-your-key-here
LLM_MODEL=poolside/laguna-m.1

# Option 2: Anthropic API key
# Get from https://console.anthropic.com/
# LLM_API_KEY=sk-ant-api03-your-key-here
# LLM_MODEL=claude-3-5-sonnet-20241022

Frontend (frontend/.env - optional):

VITE_API_URL=http://localhost:8080  # Defaults to localhost:8080

Development Mode

Set DEV_MODE=true in backend .env to bypass authentication (default user created automatically).

LLM Provider Selection

After initial setup, configure your preferred LLM provider in the admin UI:

1. Go to Admin → LLM Configuration in the web UI
2. Select your provider:
   • Poolside (Laguna models) - Recommended for best performance
   • Anthropic - Claude models
   • OpenAI - GPT models
3. Enter the model name and API key
4. For Poolside: Use poolside/laguna-m.1 or poolside/laguna-xs

Documentation

• prd.md - Full product requirements and features
• backend/CLAUDE.md - Backend development guide (testing, API, workflows)
• frontend/CLAUDE.md - Frontend development guide (API generation, stores, components)

API Documentation

Once running, visit:

• OpenAPI spec: http://localhost:8080/openapi.json
• Temporal UI: http://localhost:8233

Troubleshooting

Frontend build fails with "Cannot fetch OpenAPI spec":

• Ensure backend is running: make dev-backend
• Manually generate API client: cd frontend && npm run generate:api

Database connection errors:

• Check Docker is running: docker ps
• Restart services: make docker-down && make docker-up

Temporal workflows not running:

• Check Temporal UI: http://localhost:8233
• Ensure worker is running (part of backend)

License

MIT