Every morning I had the same routine: open Hacker News, GitHub Trending, ArXiv, TechCrunch, five AI lab blogs, and three newsletters. Thirty minutes later, I had read the same Claude announcement eight times across different sources and still missed the one ArXiv paper that actually mattered to my stack.
A conversation with a friend at Meta confirmed it: even people inside the industry cannot keep up. The firehose is useless. We need filtered, scored, summarized intelligence.
So I built it.
What ScanBrief Does
ScanBrief is an automated intelligence pipeline that runs every morning at 6:30 AM. It:
- Ingests from 20+ high-signal sources — Hacker News (API), GitHub Trending (scrape), ArXiv, OpenAI blog, Google AI blog, TechCrunch, Ars Technica, The Verge, and more coming weekly.
- Deduplicates — URL hash matching plus fuzzy title similarity. The same story appearing on HN, Reddit, Twitter, and three newsletters gets collapsed into one item.
- Scores — Each item gets a composite score (0-100) based on source authority, novelty, velocity (how fast it is spreading), and relevance to your tech stack.
- Summarizes — Claude AI generates 2-sentence summaries for the top 15 items. Enough context to decide if you need the full article.
- Delivers — Clean markdown brief to Discord, email, Slack, or via API. Read it in 5 minutes over coffee.
Why Not Just Use a Newsletter?
Newsletters are humans picking links on their schedule. That is fine for casual reading. But for builders who need to stay current on signals that affect their actual codebase:
- Newsletters miss things — human bandwidth is finite. ScanBrief scans 92+ items per run.
- No deduplication — TLDR, The Rundown, and Morning Brew all cover the same story separately. You read it three times.
- No personalization — they cannot score by what affects YOUR projects and YOUR stack.
- Fixed schedule — ScanBrief delivers when you need it, how you need it.
The Tech Stack
Deliberately simple:
- Python — ingesters, dedup engine, scorer, brief formatter
- Claude AI (Haiku) — summarization via CLI
- SQLite — item storage, dedup history, brief archive
- Discord webhooks — delivery (email and API coming in Sprint 3)
- Cron — 6:30 AM daily, runs in ~2 minutes end-to-end
Total infrastructure cost: roughly $2/month. All source APIs are free. AI summarization runs on Claude CLI. Hosting is a single Linux server.
What Makes Dedup Hard (and Valuable)
Deduplication is the killer feature and the hardest engineering problem. Stage 1 catches exact URL matches across sources. Stage 2 uses fuzzy title matching (85%+ similarity threshold) to merge "Claude 5 Released" with "Anthropic Launches Claude 5." Future sprints add semantic embeddings for complete rephrases and cross-language duplicates.
Getting this right is the difference between a useful brief and another noise source. We are conservative on thresholds — better to show a near-duplicate than to accidentally merge two different stories.
Coming Next: Project Relevance
The feature I am most excited about is project relevance mapping. ScanBrief will scan your actual codebase and tech stack, then flag when a trending signal affects YOUR projects. "New Claude version released" automatically flags every project that uses the Claude API. "GitHub ships AI code review" flags your CI/CD pipeline.
No newsletter can do this. It requires knowing what you are building.
Pricing
| Tier | Price | What You Get |
|---|---|---|
| Free | $0 | Daily email brief, top 5 items, 3 tracks |
| Pro | $9/mo | Full brief, all tracks, emerging signals, custom sources, API access |
| Team | $29/mo | Shared dashboard, team tracks, Slack/Discord webhooks, priority support |
Early subscribers get Pro features free for 3 months.
Try It
Head to scanbrief.dev and sign up for the free tier. You will get your first brief tomorrow morning.
Signal compounds. Noise compounds too. Choose which one you are stacking.
Built by Prime Automation Solutions — We build autonomous systems that run while you sleep. From network automation to AI pipelines, we help businesses eliminate manual work.