free · built in public

Your AI that reads
the internet. You just
read what matters.

inferr scrapes Hacker News and Dev.to every day, summarises each article in 3 lines, and surfaces only what's relevant to your stack — automatically.

Free during early access. No credit card.

HN · 847 pts · 94% match
Building RAG systems with pgvector and LLMs

PostgreSQL's pgvector extension enables semantic search through embeddings.

Combine with an LLM for retrieval-augmented generation in production.

Real-world example: 50K docs, <1s queries, $0.02 cost per summary.

NestJS · RAG · PostgreSQL

Daily digest

HN + Dev.to scraped every 24h. Top posts surface automatically — no manual browsing.

  • Hacker News top stories
  • Dev.to trending articles
  • Zero duplicate noise

Always consistent. Always available.

Today's feed
Rust 2024 edition: what changed
TypeScript 5.5 inference upgrades
pgvector HNSW vs IVFFlat benchmarks
HN · Dev.to · 30 new today

3-line AI summaries

gpt-4o-mini distills each article to three lines. No fluff, just the point.

  • One sentence per key insight
  • Why it matters to developers
  • Cost: $0.02 per 100 articles

Always consistent. Always available.

⚡ AI Summary
How Rust handles memory without GC
1. Ownership rules enforce memory safety at compile time.
2. Borrow checker prevents data races with zero runtime cost.
3. Drop trait replaces destructors — no GC pause, ever.

Interest-ranked feed

Set your stack once. Embeddings rank articles by relevance to you, not by recency.

  • Semantic matching via pgvector
  • Your tags → 1536-dim query vector
  • Re-ranks on every feed load

Always consistent. Always available.

#1Rust96% match
#2TypeScript91% match
#3NestJS87% match
#4pgvector82% match

Semantic chat

Ask questions across your feed. "What's new in vector databases this week?" actually works.

  • RAG over your ranked articles
  • Sources cited in every answer
  • Powered by gpt-4o-mini

Always consistent. Always available.

What's new in vector databases this week?
pgvector 0.8 shipped with HNSW improvements. Qdrant released sparse vector support. Weaviate added multi-tenancy at the cluster level.
Sources: 3 articles from your feed

Up and running in 30 seconds

01

Sign in with Google

One click. No password, no setup.

02

Set your interests

Tell us your stack. NestJS, Rust, LLMs — whatever you actually work with.

03

Feed builds itself

Articles scraped, summarised, and ranked for you every 24h.

04

Ask it anything

Use semantic chat to query your feed like a database.

NestJS · TypeScript · PostgreSQL · pgvector · Redis · gpt-4o-mini · Railway

One repo. Open source. $7/month to run. View on GitHub →

Stop losing 45 minutes to
Hacker News every morning.

Free. Takes 30 seconds to set up.